United States
Environmental Protection
Agency
Risk Reduction
Engineering Laboratory
Cincinnati OH 45268
Research and Development
EPA/600/SR-92/120 Sept. 1992
EPA Project Summary
Total Hydrocarbon Emission
Testing of Wastewater Sludge
Incinerators
John T. Chehaske, William G. DeWees, and F. Michael Lewis
The U.S. Environmental Protection
Agency (EPA) is considering a regula-
tory requirement for continuous moni-
toring of total hydrocarbon (THC) emis-
sions from all wastewater sludge incin-
erators. This study was conducted to
determine the reliability of total hydro-
carbon analyzers (THCAs) in this appli-
cation.
Continuous monitors for oxygen (O2),
carbon monoxide (CO), THC, and tem-
perature were installed at two munici-
pal wastewater sludge incinerators. The
O2 data were used to normalize the
measured THC concentrations to 7%
O2. CO was measured to determine if it
could be used as a surrogate for THC
measurements.
The two THCAs performed very well,
achieving 94% and 90% on-line avail-
ability at the two sampling sites, re-
spectively. The O2 and CO analyzers
also worked well. There were initial
problems with the sample conditioning
system that is necessary for the CO
and Os monitors, but successful opera-
tion was achieved after it was modi-
fied. The average hourly THC concen-
trations, normalized to 7% O,, were 108
and 88 parts per million by volume
(ppmv) for the two sites.
The corresponding CO concentra-
tions were 1,071 and 1,091 ppmv. The
CO and THC data showed a poor corre-
lation. Several short runs were made
under high turbulence firing conditions,
and others were made with the use of
an afterburner. These short runs ranged
from 13 min to about 5 hr and achieved
O2 normalized THC concentrations of
less than 30 ppmv.
This Project Summary was developed
by EPA's Risk Reduction Engineering
Laboratory, Cincinnati, OH, 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
The EPA Office of Water (OW) drafted
risk-based sludge regulations (for incin-
eration and a variety of other options)
under Section 405d of the Clean Water
Act. The proposed regulations were pub-
lished for public comment in the Federal
Register." Being considered for the final
regulation is a provision for continuously
monitoring THC emissions as a method of
controlling organic emissions from sludge
incineration. The monitoring would have
to demonstrate that the THC stack emis-
sions were not exceeding a concentration
limit.
To evaluate THC monitoring the EPA
Risk Reduction Engineering Laboratory
(RREL) implemented a study for OW to
operate total hydrocarbon analyzers
(THCAs) over a 3-mo period at two mu-
nicipal wastewater sludge incineration fa-
cilities. Continuous analyzers for THC, CO,
and O2 were installed and operated at the
two facilities, (Site THC-1 and Site THC-
2) both of which employed multiple-hearth
furnaces (MHFs) to incinerate wastewater
sludge. In addition, EPA requested an
evaluation of the use of these monitors to
assist with improving incinerator opera-
tion.
Fed. Reg., 54 (23): 5746-5902, Feb. 6,1989.
Printed on Recycled Paper
-------
Description of Facilities
Site THC-1 had been previously stud-
ied by the EPA. The results of that study
are contained in EPA report "Emissions of
Metals, Chromium and Nickel Species, and
Organics From Municipal Wastewater
Sludge Incinerators, Volume III: Site 6
Emission Test Report"*. Site THC-2 had
also been previously studied, but different
incinerators were tested in that EPA study.
She THC-1 is a publicly operated treat-
ment works (POTW) that processes an
average of 30 million gallons per day (mgd)
of wastewater. The sludge furnace oper-
ates 24 hr per day, 5-3/4 days per week.
Sludge is dewatered before incineration
to reduce sludge-cake water content to
70% to 75% by weight. A gravity thick-
ener is used to increase the percentage
of solids in the primary sludge, and a
flotation thickener processes the second-
ary sludge. Lime slurry and ferric chloride
solution are used to condition the sludge
drawn from a storage tank. Four recessed-
plate filter presses are used to dewater
tha conditioned sludge before transport to
the furnaces.
Site THC-1 has two identical Nichols*
eight-hearth, MHFs; only one is operated
at a time. Sludge is screw fed into the
side of Hearth #1. The air pollution control
system consists of an adjustable-throat
venturi scrubber, followed by a two-plate
impingement tray scrubber.
Tha project at Site THC-1 was con-
ducted under normal plant and incinerator
operating conditions. During the test pro-
gram, the unit operated a total of 1,690 hr
out of a total of 2,232 possible hr, repre-
senting a 76% on-line operation.
Site THC-2 is a POTW treating approxi-
mately 40 mgd of primarily domestic waste-
water. Solids processes at the facility are:
gravity thickening of chemically treated (fer-
ric chloride) primary sludge; flotation thick-
ening of waste-activated sludge; mixing/
storage of thickened sludges; chemical
conditioning/dewatering of sludge by a
combination of a vacuum filter, belt press,
and membrane-type filter press; and in-
cineration using multiple hearth incinera-
tors.
Site THC-2 has two identical MHFs,
only one of which is operated at a time.
During this study, both furnaces were used.
Tha furnaces are eight-hearth units, ap-
proximately 22 ft diameter, built in 1979
by Nichols Engineering.
tEPA 60Q,'R-32/003c. PB 92-151570/AS. March 1992.
'Mention of trado names or commercial products does
not constitute endorsement or recommendation for
Throughout most of this study, tests at
Site THC-2 were conducted under normal
plant and incinerator operating conditions.
Because of ongoing plant modifications,
however, several problems were encoun-
tered. These problems related to changes
in the dewatering system and modifica-
tions to the burner control circuit. Other
problems at Site THC-2 included the fail-
ure of an induced draft (ID) fan bearing
and a fire in the sludge conveyor and
control wiring. During the test program,
Site THC-2 operated a total of 1,533 hours
out of a possible 2,184 hours, represent-
ing a 70% on-line operation.
Procedures
Sampling sites at both facilities were
located in a transition section between the
ID fan and the outlet stack, downstream
of the scrubber and demister but upstream
of the shaft cooling air point of entry. De-
tailed descriptions of each sampling site
are presented in the report.
Continuous emission monitors (CEMs)
were used to monitor THC, CO, and O in
the sludge incinerator exhaust gases. Ex-
haust gas temperature was monitored with
a thermocouple. The analyzer and ther-
mocouple outputs were recorded on a
computer with the use of an electronic
data acquisition system. The analyzer out-
puts were also recorded on strip charts.
The THCA sampling system for both
sites consisted of a probe, a three-way
calibration valve, and a heated sample
line. The heated sample line connected
directly to the oven section of the THC as
manufactured by J.U.M. Engineering. In-
side the oven, which was maintained at
190°C, was a sample filter, a pump, a
sample flow control capillary, and a flame
ionization detector.
The Og/CO sampling system was more
complex. The initial system had a sintered
stainless steel filter attached to the probe
inlet. This system suffered repeated block-
age of the sintered stainless steel filter.
The redesigned system had a 1/4-in. OD
stainless steel probe that was open at the
inlet.
At Site THC-1, a Wilkerson compressed
air line water trap was attached to the
three-way valve, which was connected to
a Nutech heated filter box and glass filter
holder. The outlet of the filter holder was
connected to a heated sample line.
At Site THC-2, a Cast model V400G
filter jar was attached to the three-way
valve, and the heated line was attached
to the filter jar. This water knockout trap
had a 7 oz capacity whereas the one at
Site THC-1 held 32 oz. The varying com-
plexity of the two systems demonstrated
how different water knockout/filter designs
can be used, depending on the amount of
contamination in the sampled gas. The
rest of the Og/CO sample acquisition sys-
tem was the same at both sites, including
a two-pass refrigeration unit.
A thermocouple was installed in the re-
frigeration unit outlet line to monitor the
gas stream temperature. Linearized out-
put from the temperature readout was fed
to the data logger. The THCA unit at Site
THC-1 was a model VE-7AP, and the unit
at Site THC-2 was a model VE-7 with an
add-on autopurge module. The THC ana-
lyzers were calibrated by using propane
in air standards prepared by Scott Spe-
cialty Gases.
Servomex model 1420 O2 analyzers
measured the O2 content of the exhaust
gas at the THC sampling location. These
analyzers utilize the paramagnetic mea-
surement technique.
Milton Roy Corporation model 3300 ana-
lyzers measured the CO concentration in
the exhaust gases. These analyzers use
a nondispersive infrared technique to mea-
sure gas concentration. The exhaust gas
temperature was measured with an Omega
Engineering, Inc., model KQ55-14G Type
K thermocouple connected to an Omega
model TAC-30 analog converter. The ther-
mocouple produces a voltage proportional
to temperature, and the analog converter
provides electronic reference junction com-
pensation and a linearized output so that
1 millivolt equals 1 degree.
Outputs from the THC, CO, and O2 ana-
lyzers and the various thermocouples were
recorded on an Odessa model DSM 3260
data acquisition/data reduction system.
The Odessa was connected to a
Walkabout 386-SX laptop computer.
Most data processing was performed
by DEECO, Inc., as DEECO had pro-
cessed continuous monitoring data from
other EPA test sites. Data from this study
were recorded in the same format used in
previous studies so EPA could enter the
data into its THC database. THC concen-
trations, measured on a wet basis, were
converted to a dry basis.
Columns were created in the data
spreadsheets for: THC concentrations cor-
rected to 7% O2; hourly average O2 con-
centrations; hourly average CO concen-
trations with off-scale data; hourly aver-
age CO concentrations with no off-scale
data; hourly average O2-corrected THC
concentrations with off-scale data; and
hourly average O.-corrected THC concen-
trations with no off-scale data.
A comprehensive quality assurance/
quality control (QA/QC) program was
implemented for this testing project to as-
-------
sess the data quality and to establish limi-
tations on the ultimate use of the data.
The objective of this QA/QC program was
to optimize precision, accuracy, complete-
ness, comparability, and representative-
ness for each major measurement param-
eter of the test program. Data quality ob-
jectives for the THC, CO, and O2 continu-
ous monitors are presented in Table 1.
Table 1. Summary of Estimated Precision, Ac-
curacy, and Completeness Objectives
Parameter
Objective
Precision
Accuracy
Completeness
±20%'
±20%f
1 Coefficient of variation (CV) determined from
periodic analyses of a control sample, where
% CV= Stand Deviation x 100
Mean
f Relative accuracy (RA) determined from the
difference between the known concentration
of the cylinder gases and the concentrations
indicated by the CEMs, where RA = /differ-
ence/actual concentration x 1001
f Valid data percentage of total hours sampled.
Four-point calibration error (linearity)
checks were made at the beginning and
end of the study and each time a new or
modified analyzer was installed. Collected
data show that the analyzers were rou-
tinely well within the allowable ±3% of
span. The maximum error observed for
any of the linearity checks was 1.6% for
CO, 0.87% for O2, and 1.0% for THC.
Bias checks were conducted at the be-
ginning and end of the study and periodi-
cally during the study when the sampling
line was disassembled to ensure that the
sampling system was not leaking. All bias
checks were below the allowable ±5% of
the control gas value. The maximum bias
check error observed was 1.6%.
Zero and span drift checks were per-
formed throughout the study. The allow-
able zero and span drift error, calculated
as a percentage of the span, was ±5%.
The maximum zero and span drift ob-
served were 0.6% and 2.0%, respectively.
Because different coefficient of varia-
tion had to be calculated for each span
gas used during the study, there was more
than one value for each analyzer. All of
the coefficient of variation results were
well below the maximum allowable ±20%.
The relative accuracy of the continuous
monitors was determined from the analy-
sis of special audit gases. The results
were all well below the ±20% maximum
allowable. The largest relative accuracy
error was 2.35%.
Results
At Site THC-1, data collection began on
June 8, 1991, and ended on September
9, 1991. At Site THC-2, monitoring ran
from June 11, 1991, to September 10,
1991.
During the study, the various analyzers
worked well with a minimum of mainte-
nance under a wide range of operating
conditions. Concentrations of THC and CO
were, however, often much higher than
anticipated.
Significant problems occurred at the be-
ginning of the study with the sample con-
ditioning system for the O2 and CO ana-
lyzers. The initial sample conditioning sys-
tem design was prone to plugging, but
after the system was redesigned, satisfac-
tory operation was achieved.
In future applications, it appears that
problems associated with particulate and
liquid water removal can be solved with
relatively simple water knockouts and fil-
ters. Water vapor removal, however, re-
quires more complicated mechanical
equipment that will increase maintenance
requirements by plant personnel.
The EPA criteria for determining THCA
analyzer reliability was on-line availability
of the analyzers. These results are pre-
sented in Table 2.
Table 2. THCA On-Line Availability
Site
THC-1
THC-2
Furnace
Operating
Hours
1,681
1,508
THCA
Operating
Hours
1,579
1,355
THCA
Availability
(%)
94
90
The target of having the THCAs operat-
ing at least 90% of the time that the plant
was operating was achieved at both sites.
At Site THC-1, 94% availability was
achieved and at Site THC-2, 90%.
A THCA failed to reignite on nine occa-
sions after an automatic backpurge cycle.
Six of these occurrences, at Site THC-1
during August, resulted in 53 hr of lost
data. The problem was corrected by ad-
justing the flame-out potentiometer inside
the analyzer. The THCA failed to reignite
three times at Site THC-2, and 184 hr of
data were lost. These failures occurred in
July, August, and September.
Had the flameout been noticed sooner
during routine inspection, the analyzer
could have been repaired more quickly,
and Site THC-2's availability would have
been considerably better. No other repairs
or maintenance had to be performed other
than the flame-out potentiometer adjust-
ment and a sample line clean out. A total
of 3 hr and 15 min were spent maintaining
the two THCAs.
THC concentrations were initially sig-
nificantly higher than had been anticipated
and resulted in stack gas concentrations
that exceeded the designated range of
the THCAs. Operating ranges of the
THCAs were subsequently expanded to
compensate for this problem. THC con-
centrations were normalized to 7% O2.
Daily averages were calculated for those
days that had 13 or more hours of data.
Neither plant achieved a O2-normalized
THC concentration of 30 ppmv or less.
Most were considerably higher. At Site
THC-1, concentrations of less than 30
ppmv or less were achieved for only 100
hr and at Site THC-2 for 51 hr.
The overall average hourly concentra-
tion at THC-1 was £108 ppmv and at
THC-2, >88 ppmv. The average daily con-
centration: (1) for days with all data on
scale was 81 ppmv for THC-1 (10 days)
and 74 ppmv for THC-2 (29 days); (2) for
days with off-scale data was >117 ppmv
for THC-1 (25 days) and >118 ppmv for
THC-2 (5 days); and (3) as an overall
average was >107 ppmv for THC-1 and
>81 ppmv for THC-2. The maximum daily
concentration for THC-1 was 5272 ppmv
and for THC-2 >188; the minimum daily
concentration for THC-1 was 40 ppmv and
for THC-2, 44. For periods with off-scale
values, the maximum scale reading was
used for calculation of these averages,
yielding equal-to-or-greater-than values.
The CO monitors worked very well dur-
ing the study, and there was one serious
problem and two minor problems with the
O monitors. The O2 monitor at Site THC-
1 had a loose output connector that caused
lost data in July, and in September, the
analyzer failed completely. A replacement
analyzer was obtained, and the failed unit
was returned for repair.
Another minor problem was encountered
with the Site THC-2 O2 monitor, when the
dumbbell inside the unit became stuck.
The dumbell deflects in a varing magnetic
field to provide a reading proportional to
O2 concentration. Tapping the side of the
case released the stuck dumbell, and no
other problems occurred.
The CO data from both sites had fre-
quent off-scale excursions. The concen-
trations were higher than had been en-
countered on previous EPA tests. Thus,
expanded-range analyzers were installed
at Site THC-1 on August 15 and at Site
THC-2 on August 5. Even with the higher
range, both sites still had a number of off-
scale excursions. The overall average
hourly concentration for the THC-1 was
1,071 ppmv and for THC-2, 1,091 ppmv.
-------
In comments to EPA, CO was suggested
as a surrogate for THC measurements.
This led to a second objective of the
proJQCt-^o determine correlations between
THC and CO concentrations. Data from
August were chosen for these analyses
because both the analyzers and the fur-
naces performed best during that time pe-
riod. For meaningful analysis, it is also
important that the correlation be deter-
mined for values near the expected emis-
sion standard. Results of the linear re-
gression analyses for data ranging from
10 to 50 ppmv are shown in Table 3. The
correlations developed for both sites do
not appear to support the use of CO con-
centrations for estimating THC concentra-
tions.
The use of breaching temperature as a
possible surrogate for THC concentrations
was also suggested. These correlation
analyses were also performed with Au-
gust data. Results of the analyses for Sites
THC-1 and THC-2 (Table 4) do not reflect
a strong correlation between THC and
breaching temperature.
Several operations test runs were con-
ducted at Site THC-2 during the weeks of
August 5 and August 12, 1991, to deter-
mine if tower THC emissions could be
obtained with relatively minor changes in
furnace hardware or operating conditions,
or both. Two different scenarios were tried.
One involved inducing more turbulence
within the hearths with special air-fuel mix-
ing valves (i.e., a high turbulence [HT]
run). In addition to this analysis, several
runs were made using the afterburner to
evaluate THC and CO reduction at differ-
ent afterburner temperatures.
The HT run was achieved during a 5-hr
period between August 8 and August 9,
1991. THC concentrations were shown to
vary between 10 and 25 ppmv normalized
to 7% O2 during this 5-hr test, with the
exceptions of a spike condition resulting
in a value of approximately 65 ppmv (nor-
malized). The spike condition was caused
by a momentary stoppage in the sludge
feed.
A second series of runs, on August 14,
1991, evaluated the use of an afterburner
to achieve desired emission limits. These
tests were at an afterburner temperature
of 1400°F. During this series of runs, THC
emissions decreased from an observed
value of approximately 50 ppmv before
starting afterburner to approximately 5 to
10 ppmv whenever the afterburner was
on.
On August 16,1991, a second analysis
of afterburner effectiveness was conducted
at temperatures between 1100°F and
1400°F. These test runs showed an in-
Table 3. Correlation of THC and CO
Site
THC-1
THC-2
Linear Regression
Equation
CO = (20.4)(THC) + 247
CO = (22.0)(THC) -i- 232
Number of
Observations
3,028
8,190
R*
0.525
0.455
Table 4. Correlation of THC and Breaching Temperature
Site
THC-1
THC-2
Linear Regression
Equation
Temp. = (-1.102)(THC) + 929
Temp. = (-1.316)(THC) + 867
Number of
Observations
3,028
8,190
R2
0.012
0.085
verse relationship of afterburner tempera-
ture to THC emissions. Under these re-
duced afterburner temperatures, emissions
varied between 7 to 21 ppmv.
Conclusions
In terms of equipment operation, the
two J.U.M. Engineering THCAs worked
reliably at both MHF wastewater sludge
incinerators, ft is believed that the suc-
cess of these analyzers was significantly
affected by the unique sample filtering and
backpurging system used by J.U.M. The
CO analyzers also performed very well.
No downtime was caused by analyzer fail-
ure; rather all downtime was caused by
failures of the sample acquisition/condi-
tioning system, which was external to the
analyzers. The Servomex O2 analyzers
performed well, despite two analyzer-re-
lated problems.
Based on the data collected in this study,
there was poor correlation between THC
and CO concentrations, and there was
essentially no correlation between THC
concentration and breaching temperature.
As indicated in this summary, sample con-
ditioning systems are critical to acquisition
of reliable CO and O data when using
extractive monitors. If not properly de-
signed, or maintained, or both, such
sample conditioning systems can cause
significant operating problems. For O2
monitors, the sample conditioning system
can be eliminated by using an in-situ moni-
tor rather than an extractive monitor.
On a limited basis, the study examined
operating the incinerator afterburner at
lower temperatures and also examined
two potentially less expensive process
modifications (that might offer an alterna-
tive to afterburners). External afterburners
are effective in reducing CO and THC
emissions, and it appeared that emissions
less than 30 ppmv may be achieved at
temperatures less than 1,400°F. Only a
short demonstration of this was possible
during this study, however. A limited test
program also demonstrated that lower THC
emissions may be achieved by increasing
the gas phase turbulence within the
hearths while simultaneously raising the
exhaust temperature by firing most of the
auxiliary fuel above the combustion hearth.
It was beyond the scope of this study,
however, to demonstrate that these tech-
niques are viable for long-term operation.
The test program reverified that an MHF
is inherently unstable. The extremely wide
variation in CO and THC emissions shown
in this study indicates that substantial
changes in operating procedures probably
will be required to comply with the new
regulations. For many plants, equipment
modifications may also be necessary.
Acknowledgments
The full reports were submitted in fulfill-
ment of Contract No. 68-CO-0027 by
James M. Montgomery Engineers under
subcontract to'Pacific Environmental Ser-
vices, Inc., DEECO, Inc., and F. Michael
Lewis, Consultant, under the sponsorship
of the U.S. Environmental Protection
Agency. Papers containing data summa-
ries have also been presented at the 1992
annual meeting of the Air and Waste Man-
agement Association and the 1992 sludge
specialty conference of the Water Envi-
ronment Federation.
•U.S. Government Printing Office: 1992— 648-080/60065
-------
-------
John T. Chehaske is with Pacific Environmental Services, Inc., Herndon, VA;
William G. DeWees is with DEECO, Inc., Gary, NC; and F. Michael Lewis is a
private consultant In El Segundo, CA.
Harry E Bosttan is the EPA Project Officer (see below).
The complete report, entitled "Total Hydrocarbon Emission Testing of Wastewater
Sludge Incinerators," (Order No. PB92-197086/AS; Cost: $19.00; subject to
change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springf^, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Cincinnati OH, 45268
United States
Environmental Protection Agency
Center for Environmental Research Information
Cincinnati, OH 45268
Official Business
Penalty for Private Use
$300
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
EPA/600/SR-92/120
------- |