-------
2.0 BACKGROUND
On July 7, 1987, EPA published an Advance Notice of Proposed Rulemaking
(ANPR) which announced EPA's intent to develop new source performance
standards (NSPS) for new MWCs, and emission guidelines for existing MWCs under
the authority of Section 111 of the Clean Air Act.2 In conjunction with the
ANPR. EPA issued operational guidance for new mass burn waterwall, modular
starved air. and refuse-derived-fuel fired MWCs. The operational guidance was
intended to serve as an interim tool to be used in Best Available Control
Technology (BACT) determinations for siting new MWCs under the Prevention of
Significant Deterioration (PSD) provisions of the Clean Air Act. The guidance
specified that combustion controls are a necessary part of BACT for new MWCs.
The background information that led to the MWC regulatory decision was
compiled and published in a Report to Congress.3 As part of this effort.
preliminary recommendations were made defining good combustion practices for
new mass burn waterwall, modular starved air, and RDF fired MWCs.4 Good
combustion practices are expected to minimize emission of organics from MWC
systems. The original recommendations included three elements:
• Design
• Operation/Control
• Verification
The requirements to satisfy these elements are:
• MWCs must be designed in a manner that minimizes air emissions.
• MWCs must be operated within an envelope dictated by the design of
the combustion system, and controls must be in place to prevent
operation outside of the established operating envelope.
• The performance of the combustion system must be verified by way
of compliance testing and through continuous monitoring of key
design and operating parameters, such as combustion air flows, gas
temperatures. CO flue gas concentrations, and 02 flue gas
concentrations.
2-1
-------
The revised recommendations for good combustion practices used in this report
are presented in Table 2-1. The revised recommendations are similar to those
provided in the Report to Congress with the exception of changes in some of
the CO emission limits and the addition of a recommendation on PM control
device operating temperature to address low temperature formation of CDD/CDF.
The good combustion practices defined in this report are designed to:
(1) maximize in-furnace destruction of organic compounds, and (2) minimize
conditions that lead to low temperature formation of CDD/CDF. Conditions
within the combustion process that satisfy the first goal includes:
• Mixing of fuel and air to minimize the existence of long-lived.
fuel-rich pockets of combustion products.
• Attainment of sufficiently high temperatures in the presence of
oxygen for the destruction of hydrocarbon species.
• Prevention of quench zones or low temperature pathways that will
allow partially reacted fuel (solid or gaseous) to exit the
combustion chamber.
All of these conditions are interrelated; successful destruction of trace
organic species requires that all three conditions be satisfied in the MWC
system. Mixing is not sufficient unless it is achieved at temperatures that
ensure thermal destruction of organic compounds. Completion of the mixing
process at adequate destruction temperatures prevents escape of combustibles
through low temperature pathways. Despite the continuing advancements made in
combustion control, perfect mixing will never be achieved in a combustion
system, whether conventional or waste fired. As a result, zero organic
emissions will not occur. The goal of good combustion practice is to provide
the conditions that will minimize air emissions of concern.
One important component which was not explicitly included in the
original recommendations addresses the potential for low temperature formation
of CDD/CDF. These formation phenomena have been measured at several full
scale MWCs. including those at Prince Edward Island; Pittsfield. MA; North
Andover. MA; and Pinellas County. FL.5.6.7.8
2-2
-------
TABLE 2-1. GOOD COMBUSTION PRACTICES FOR MINIMIZING TRACE ORGANIC
EMISSIONS FROM MUNICIPAL WASTE COMBUSTORS
Temperature at fully
mixed height
Underfire air control
Overfire air capacity (not
an operating requirement)
Overfire air injector
design
Auxiliary fuel capacity
Downstream flue gas
temperature
1800°F (982°C)
At least 4 separately adjustable plenums.
One each under the drying and burnout zones
and at least two separately adjustable
plenums under the burning zone (MB/WW). As
required to provide uniform bed burning
stoichiometry (RDF)
40% of total air (MB/WW. RDF)
80% of total air (MOD/SA)
That required for penetration and coverage
of furnace cross-section
That required to meet start-up temperature
and 1800°F (982°C) criteria under part-load
conditions
<450°F «232°C) at PM control device
inlet
OPERATION/CONTROL
Excess air
Turndown restrictions
Start-up procedures
Use of auxiliary fuel
6-12% oxygen in flue gas (dry basis) (MB/WW
and MOD/SA). 3-9% oxygen in flue gas (dry
basis) (RDF)
80-110% of design - lower limit may be
extended with verification tests
On auxiliary fuel to design temperature
On prolonged high CO or low furnace
temperature
VERIFICATION
Oxygen in flue gas
CO in flue gas
Furnace temperature
Adequate air distribution
Downstream flue gas
temperature
Monitor
Monitor - 50 ppm on 4 hour average.
corrected to 7% 02 (MB/WW and MOD/SA). 100
ppm at 7% 02 (RDF)
Monitor - minimum of 1800°F (982°C) (mean)
at fully mixed height across furnace
Verification Tests
Monitor - <450°F «232°C) at PM control
device inlet
MB/WW - mass burn waterwall
MOD/SA - modular starved air
2-3
-------
The discovery of CDO/CDF formation in full scale MWCs has prompted
research in the laboratory to identify the parameters controlling low
temperature CDD/CDF formation reactions. Bench scale experiments indicate
that, under excess air conditions. CDD/CDF formation occurs on the surface of
fly ash at temperatures ranging from approximately 200 to 400°C (392 to
752°F), with the maximum formation occurring near 300°C (572°F). 9 Conversely,
research results have indicated that, when the same experiments were performed
in an oxygen-deficient atmosphere, dechlorination of CDD/CDF compounds
occurred.io The current thinking regarding these findings is that the
formation process may involve catalytic reactions of organic precursor
compounds with particulates containing metallic species such as copper
chloride (CuCl?). The bench scale studies indicate that the rate of CDD/CDF
formation and/or chlorination is affected by a number of parameters, including
temperature, residence time, catalyst effects, carbon content, oxygen
concentration, and moisture. Results from these experiments provide
information which can be transferred to full scale MWCs in order to develop
control strategies for minimizing CDD/CDF formation.
Although many strategies for minimizing the reactions (e.g., catalyst
poisons) remain to be investigated, it appears at this time that an initial
control strategy is to minimize the particulate matter concentration and the
flue gas residence time at temperatures were the rate of CDD/CDF formation is
highest. If organic precursor materials leaving the combustor are minimized
and if flue gas retention times and PM concentrations can be minimized in the
200-400°C (392-752°C) temperature range, it appears that the formation process
can be minimized. Many existing MWCs currently operate flue gas cleaning
equipment (ESPs) in this temperature window. The increased gas residence time
and PM concentrations which occur in the ESP may be the primary cause of
CDD/CDF formation, leading to increased emissions in the stack. Recent data
from a full scale MWC confirm that high efficiency ESPs operating at
temperatures below 250°C actually provide significant CDD/CDF removal.11 Based
on these considerations, a new component of good combustion practices was
developed. The recommendation is to maintain PM control device inlet gas
temperatures below 232°C (450°F).
2-4
-------
3.0 MODEL PLANT PERFORMANCE ESTIMATES - MASS BURN COMBUSTORS
The following sections discuss the data and provide the rationale used
to establish the baseline emission values in Table 1-2. The subsections are
organized according to model plant combustion technology. The data used to
establish baseline emissions for each model plant are compiled from emission
tests performed at some of the newest plants in the existing MWC population.
The emission tests can be separated into three distinct categories:
1. Compliance tests with sampling performed at the stack, downstream
of flue gas cleaning equipment. In most cases these data were
generated under optimal operating conditions at or near design
steam load. In many cases process data such as temperatures and
airflows were not recorded during testing.
2. Compliance tests with sampling performed concurrently at the inlet
and outlet of the flue gas cleaning equipment. In most cases.
limited process data were recorded, and the combustor operated at
or near design steam load.
3. Parametric tests involving multi-point sampling under a variety of
combustor and/or flue gas cleaning device operating conditions.
Process data are usually well documented in these test reports.
The emissions data used in this analysis are presented in both tabular
and graphical form. The data tables present multiple run averages reported
for each test facility. Data measured in a parametric test are averaged and
presented separately for each parametric operating condition. Combustor
design and operating data are also included in the data summary tables. The
data graphs present the emission levels for each sampling run, along with an
average value for each testing condition.
Test reports that include emissions measured upstream of flue gas
cleaning equipment provide the best data to evaluate combustion conditions.
As a result, emission tests in categories 2 and 3 are the primary focus of
this analysis. In some cases, data measured in the stack also provide
information related to combustion conditions. When these data offer some
insight into combustion conditions experienced during testing, they also
3-1
-------
provide insight into combustion conditions experienced during testing; they
are also included in the discussion of baseline emissions.
3.1 Conventional Mass Burn Waterwall MWCs
There are three conventional mass burn waterwall models included in this
analysis. An inventory of planned and projected MWCs indicates that the
majority of new facilities subject to the lll(b) NSPS will use mass burn
technology.12 The three model plants are representative of the projected
population of MWCs. comprising two combustors at 100 tpd (91 Mg/day) each; two
combustors at 400 tpd (364 Mg/day) each; and three combustors at 750 tpd (681
Mg/day) each.
3.1.1 Emissions Data for Existing Facilities
The baseline emissions are established based on review of available
emissions data from some of the newest mass burn waterwall MWCs in the
existing population. These units are considered to be representative of the
planned and projected MWC population as of September 1988. Data are available
for facilities using Von Roll. Martin. Detroit Stoker, and Riley Takuma
technologies. These four system designs comprise more than 85 percent of the
existing mass burn waterwall MWCs, and each manufacturer is expected to
continue to be represented in the new MWC population. Descriptions of generic
design and operating features associated with each of the system designs are
included in EPA's Report to Congress on Municipal Waste Combustion.4
Table 3-1 presents a summary of emissions data for the facilities
included in this analysis. Also included in Table 3-1 is information
detailing the design and operation of each facility relative to the good
combustion practice criteria which was developed for mass burn waterwall MWCs.
When available, combustor operating conditions are presented as reported
during the actual testing period. In cases where process operating data are
not available, information has been supplied for a facility as reported in the
Clean Air Act Section 114 questionnaire responses which were submitted to EPA
by each existing facility.
3-2
-------
TABLE 3-1. LARGE MASS BURN WATERWALL MWCS - PERFORMANCE ASSESSMENT
PAGE 1 OF 4
FACILITY
NUMBER OF UNITS - Flue gas cleaning equipment (FGC)
UNIT SIZE, tpd (Mg/day)
UNCONTROLLED EMISSIONS
CDD/CDF (ng/dscm)
CO (ppmv)
PM (mg/dscm)
CONTROLLED EMISSIONS
CDD/CDF (ng/dscm)
CO (ppmv)
COMBUSTION PARAMETERS
GOOD COMBUSTION
PRACTICE RECOMMENDATIONS
Mi 11 bury. MA
2 - SD/ESP
750 (682)
170
38
NA
59.2
FACILITY DESIGN
AND OPERATING CONDITIONS
Temperature at fully
mixed height
Underfire air
Overfire air capacity
(not an operating
requirement)
Overfire air injector
design
Auxiliary fuel capacity
Exit gas temperature
OPERATION
Excess air
Turndown
Overfire air
Start-up procedures
Auxiliary fuel use
VERIFICATION
02 levels
CO
Temperature
Air distribution
Exit gas temperature
1800°F (982°C) mean
At least 4 plenums along
grate length
40% total air
Complete penetration/
coverage
As required to achieve
temperature limits
during start-up
<450°F (232°C) at PM
control device inlet
6-12% 02 (dry)
80-110% design load
Penetration and coverage
of furnace cross section
Auxiliary fuel to design
temperature
High CO. low temp:
start-up/shutdown
Monitor
Monitor «50 ppm at 7% 02)
Monitor
Monitor
Monitor
1500°F (816°C) at
superheater inlet
5 plenums along
grate length
At least 60% total air
3 rows (2 front. 1 rear)
Gas - 40% load
435°F (224°C)
10.2% 02
Baseloaded - 100* ±3%;
66X minimum
40-50% total air
Gas - 1500°F (816°C) at
superheater inlet
Start-up/shutdown
Yes
Yes
Superheater inlet/
outlet
OFA, UFA pressures
Yes
3-3
-------
TABLE 3-1. LARGE MASS BURN WATERWALL MWCS - PERFORMANCE ASSESSMENT
PAGE 2 OF 4
FACILITY
NUMBER OF UNITS - FGC
UNIT SIZE, tpd (Mg/day)
UNCONTROLLED EMISSIONS
CDD/CDF (ng/dscm)
CO (ppmv)
PM (mg/dscm)
CONTROLLED EMISSIONS
CDD/CDF (ng/dscm)
CO (ppmv)
COMBUSTION PARAMETERS
Pinellas County. FL
3 - ESP
1050 (954)
69
4
225
132
FACILITY DESIGN
AND OPERATING CONDITIONS
Tulsa. OK
3 - ESP
375 (341)
36
22
FACILITY DESIGN
AND OPERATING CONDITIONS
Temperature at fully
mixed height
Underfire air
Overfire air capacity
(not an operating
requirement)
Overfire air injector
design
Auxiliary fuel capacity
Exit gas temperature
OPERATION
Excess air
Turndown
Overfire air
Start-up procedures
Auxiliary fuel use
VERIFICATION
02 levels
CO
Temperature
Air distribution
Exit gas temperature
1700°F (927°C)
5 plenums along
grate length
At least 25% total air
2 rows (1 front, 1 rear)
None
450-550°F (232-288°C)
8-10% 02 at full load
9-11* 02 at minimum load
70-90% design load
25% of total air
No auxiliary fuel
None
Yes
No
Furnace roof, super-
heater inlet/outlet
OFA pressure, UFA
damper settings
Yes
1400-1600°F (760-872°C)
at superheater inlet
5 plenums per grate
run
At least 40% total air
NA
None
375-515'F (191-263°C)
7-12% 02
72-100% load
20-40% total air
No auxiliary fuel
None
Yes
Yes
NA
OFA. UFA pressures
Yes
3-4
-------
TABLE 3-1. LARGE MASS BURN WATERWALL MWCS - PERFORMANCE ASSESSMENT
PAGE 3 OF 4
FACILITY
NUMBER OF UNITS - FGC
UNIT SIZE, tpd (Mg/day)
UNCONTROLLED EMISSIONS
CDD/CDF (ng/dscm)
CO (ppmv)
PM (mg/dscm)
CONTROLLED EMISSIONS
CDD/CDF (ng/dscm)
CO (ppmv)
COMBUSTION PARAMETERS
Marion County. OR
2 - SD/FF
275 (250)
43
18
205
1.39
FACILITY DESIGN
AND OPERATING CONDITIONS
Alexandria, VA
3 - FI/ESP
375 (341)
53
18
FACILITY DESIGN
AND OPERATING CONDITIONS
fiESJLGl
Temperature at fully
mixed height
Underfire air
Overfire air capacity
(not an operating
requirement)
Overfire air injector
design
Auxiliary fuel capacity
Exit gas temperature
OPERATION
Excess air
Turndown
Overfire air
Start-up procedures
Auxiliary fuel use
VERIFICATION
02 levels
CO
Temperature
Air distribution
Exit gas temperature
1400-1600°F (760-872°C)
at superheater inlet
5 plenums per grate run
At least 40% total air
3 rows
Gas - 30% load
392°F (200°C)
7-12% 02
75-105% design load
20-40% total air
Gas to 1800°F (982°C)
Start-up/shutdown
Yes
No
Middle and top of
furnace
OFA. UFA pressures
Yes
1400-1600°F (760-872°C)
at superheater inlet
5 plenums along grate
length
At least 40% total air
2 rows
Oil - 25% thermal load
375-505°F (191-263°C)
7-12% 02
80-100% load
20-40% total air
On oil to 1400°F (760°C)
Start-up/shutdown
Yes
Yes
Furnace exit
(superheater inlet)
OFA. UFA pressures
Yes
3-5
-------
TABLE 3-1. LARGE MASS BURN WATERWALL MWCS - PERFORMANCE ASSESSMENT
PAGE 4 OF 4
FACILITY
NUMBER OF UNITS - FGC
UNIT SIZE, tpd (Mg/day)
UNCONTROLLED EMISSIONS
COD/CDF (ng/dscm)
CO (ppmv)
PM (mg/dscm)
CONTROLLED EMISSIONS
CDD/COF (ng/dscm)
CO (ppmv)
COMBUSTION PARAMETERS
Commerce. CA
1 - SD/FF
350 (318)
27
4620
1.70
16
FACILITY DESIGN
AND OPERATING CONDITIONS
Olmstead County. MN
2 - ESP
100 (91)
31-54
FACILITY DESIGN
AND OPERATING CONDITIONS
Temperature at fully
mixed height
Underfire air
Overfire air capacity
(not an operating
requirement)
Overfire air injector
design
Auxiliary fuel capacity
Exit gas temperature
OPERATION
Excess air
Turndown
Overfire air
Start-up procedures
Auxiliary fuel use
VERIFICATION
02 levels
CO
Temperature
Air distribution
Exit gas temperature
1700°F (926°C)
at superheater inlet
6 plenums (2 per grate
length)
40% total air
2 front, 2 rear. 1 side
Gas - 100% load
480°F (249°C)
10% 02 ±2%
70-101% design load
20-40% of total air
On gas
Start-up/shutdown
Yes
Yes
Yes
OFA. UFA pressures
Yes
1700°F (926°C)
furnace gas exit
3 plenums along
grate length
At least 35% total air
3 rows
Gas - 10% load
425°F (218°C)
7% 02
60-100% load
35% total air
On gas
Start-up/shutdown
Yes
Yes
Yes
OFA. UFA pressures
Yes
3-6
-------
3.1.1.1 Millburv. Massachusetts
Wheelabrator Environmental Systems is the U.S. licensee of Von Roll
technology. One of the newest Wheelabrator plants in operation is the
Millbury, MA Resource Recovery Facility, which includes two 750 tpd (682
Mg/day) combustors. Each unit is equipped with a spray dryer and an ESP. The
facility began operating in 1987 and underwent compliance testing in early
1988. In addition to performing stack testing for compliance purposes,
emissions were measured at the spray dryer inlet location. Five CDD/CDF
emission samples were gathered at the spray dryer inlet location at Unit #2.
and average emissions were 170 ng/dscm CDD/CDF.*3 Individual runs ranged from
140 to 210 ng/dscm. Average CO emissions were 38 ppmv (4-hour average) during
the five test runs. The average gas temperature at the inlet sampling
location ranged from 429 to 442°F (221 to 228°C) during the five runs.
An assessment of the combustor design at Millbury indicates that the
majority of design elements are in place to provide good combustion. Furnace
temperatures are measured at the inlet and outlet of the superheater. The
thermocouple at the superheater inlet location is approximately 35 feet (10.7
m) above the last point of overfire air injection, and a typical operating
temperature at this location is 1500°F (816°C). Millbury has five
individually controllable underfire air plenums along the length of the
reciprocating grates. One design feature at Millbury that differs from older
Von Roll systems is the overfire air capacity. The overfire air system has
the capacity to supply 60 percent of total combustion air. The Millbury units
operate with 40-50 percent of total air supplied as overfire air. Many Von
Roll facilities constructed prior to Millbury operate with 30-40 percent of
total airflow as overfire.
The majority of operation/control and verification elements representing
good combustion practice are also in place at Millbury. The units typically
operate at full capacity generating electricity, so low load operation is
infrequent. All the recommended monitoring procedures are in place at
Millbury.
3.1.1.2 Pinellas County. Florida
The Pinellas County MWC consists of three 1050 tpd (954 Mg/day)
combustors, with Martin stokers and three-field ESPs. The #1 and #2 units
3-7
-------
started up in 1983. and #3 commenced operation in 1986. Six ESP inlet/outlet
CDD/CDF emission tests were conducted at Unit #3 in February and March 1987.
Average inlet emissions were 69 ng/dscm and average CDD/CDF emissions in the
stack were 132 ng/dscm.8 Average flue gas temperatures at the ESP inlet
location ranged from 523 to 553°F (273 to 289°C). Average PM and CO values
measured at the ESP inlet were 0.98 gr/dscf (225 mg/dscm) and 4 ppmv,
respectively. The CO emissions were measured concurrently with CDD/CDF
testing, which was 3 hours' duration. Boiler #3 operated between 88 percent
and 91 percent rated capacity during the six test runs, and QZ concentrations
varied from 6.9 to 7.7 percent (wet basis). Average furnace temperatures
reported during testing varied from 1824°F (996°C) to 1923°F (1051°C).
measured in the upper furnace. Underfire and overfire air plenum pressures
were recorded and were fairly consistent during all of the runs. Actual
airflow splits are not available.
The design of the Pinellas County plant meets the majority of criteria
required for good combustion. However. Pinellas County does not have
auxiliary fuel burners. In addition, the normal ESP operating temperature is
approximately 500°F (260°C). The ESP temperature is assumed to have
contributed to the increased CDD/CDF concentrations measured at the ESP outlet
location. Based on the emission test results (low organics and CO levels), it
is concluded that the unit achieves good mixing. The three units at Pinellas
County are operated on a manual combustion control scheme, with the exception
that steam production rates are automatically controlled. The majority of
mass burn waterwall MWcs are equipped with fully automatic combustion
controls. A manual control scheme may allow greater potential for combustion
upsets to occur. With the exception that the units do not monitor CO
continuously, Pinellas County has all of the verification measures in place to
ensure good combustion practices are maintained.
3.1.1.3 Tulsa. Oklahoma
Emissions data from two other facilities using Martin designs were also
used to establish baseline emission factors for the model plant. The first of
these is the Tulsa, OK facility, which consists of three 375 tpd (341 Mg/day)
combustor units. The facility began operating in 1986. Each of the units is
equipped with an ESP. The ESP inlet gas temperature typically varies from
375 to 515°F (191 to 268°C). Available emissions data gathered for compliance
purposes indicate average CDD/CDF emissions of 36 ng/dscm at the stack
3-8
-------
location.14 The flue gas temperature during testing was not included in the
report. Emissions of CO averaged 22 ppmv at Unit #1 and 27 ppmv at Unit #2.
The CO data were gathered separately from the CDD/CDF data, and are presented
as 1-hour averages. There are no process data available in the test report to
use in evaluating the combustor operating conditions. With the exception that
Tulsa does not have auxiliary fuel, all of the requirements of good combustion
are assumed to be achieved at this facility.
3.1.1.4 Marion County. Oregon
The Marion County, OR MWC consists of two 275 tpd (250 Mg/day) Martin
combustors equipped with spray dryers and fabric filters. The units commenced
operation in 1986. During stack compliance testing in September 1986. EPA
performed three sampling runs at the boiler outlet (spray dryer inlet)
location. Two of the three sampling runs were invalidated, but the results of
the one successful run indicated an inlet CDD/CDF emission rate of 43 ng/dscm.
In addition, inlet particulate emissions were 0.89 gr/dscf (2050 mg/dscm). and
CO emissions were 18 ppmv (4-hour average).15
Process data were recorded during the compliance test at Marion County.
The steam load was 97 percent of design load during CDD/CDF testing. Gas
temperatures measured in the middle of the first furnace pass averaged 1741°F
(949°C), and the average economizer outlet temperature was 392°F (200CC).
Average exhaust gas oxygen concentrations were 9.5 percent, and the estimated
overfire/underfire air ratio was 25/75. The Marion County units are equipped
with auxiliary fuel burners that can provide 30 percent of thermal load. The
units do not have continuous CO monitors.
EPA gathered an additional 14 unabated CDD/CDF samples at Marion County
in February 1987. During all of the sampling runs, the boiler was operated at
normal, full load conditions. Analysis was completed on seven of the runs.
Four of the seven samples had acceptable spike recoveries and were full
traverse samples. Three of the seven runs either were single point samples or
were invalidated due to poor recoveries. The CDD/CDF values from the valid
test runs ranged from 56 to 116 ng/dscm. with an average value of 99 ng/dscm.is
3-9
-------
3.1.1.5 Alexandria. Virginia
The third new Martin facility. Alexandria, VA. consists of three 375 tpd
(341 Mg/day) units that began operating in 1987. The system design includes
in-furnace lime injection for acid gas control, and ESPs for PM control.
Results of compliance testing performed at Unit #1 in December 1987 indicated
a three-run average of 53 ng/dscm CDD/CDF and 18 ppmv of CO (3-hour average)
in the stack.17 Limited process data are included in the compliance test
report. The boiler reportedly operated at between 98 and 99 percent design
steam load during the three runs, and average 62 concentrations were 9.5
percent. The furnace temperature was measured at an unspecified furnace exit
location with an unshielded thermocouple. The average temperature during the
three runs was 1651°F (899°C), 1664°F (907°C). and 1642°F (894°C). The test
report authors state that the measurement method "should be regarded as being
relatively accurate; i.e., lower than the actual temperature by approximately
150-200°F, but precise." The average stack temperature was reported to be
342°F (172°C) while sampling, so it is judged that the ESP temperature was
below 450°F (232°C). Although the existing data base does not provide a basis
for estimating the effect of dry lime furnace injection on CDD/CDF. the
emission levels are typical of those measured at other Martin systems, that do
not use acid gas controls.
The facility is judged to satisfy the majority of criteria included in
the good combustion practice recommendations. The gas temperatures at the
superheater inlet are reported to vary from 1400°F (760°C) to 1600°F (871°C).
The plant has an auxiliary fuel source (oil), and the firing capacity is 25
percent of boiler load.
3.1.1.6 Commerce. California
The Commerce. CA. MWC consists of one 350 tpd (318 Mg/day) unit with
Detroit Stoker grates and a Foster Wheeler boiler. The unit is equipped with
a spray dryer/fabric filter. Commerce was also the first MWC in the U.S. to
use thermal de-NOx controls. The facility underwent an emissions test in 1987
for compliance purposes. The CDD/CDF emissions data were measured according
to the draft California Air Resources Board (CARB) Modified Method 5 (semi-
VOST) protocol. Two test runs were conducted at the stack while burning the
largely commercial waste normally received at the facility. A third test run
3-10
-------
was conducted with simultaneous measurement at the boiler exit and stack while
burning a residential refuse brought in from Long Beach. CA. specifically for
the test. Steam load was reduced from full load to 80 percent of capacity
during the inlet/outlet sampling. During the one test run that was conducted
at the spray dryer inlet. 27 ng/dscm CDD/CDF was measured. Inlet PM and CO
emissions were 1.56 gr/dscf (3590 mg/dscm) and 16 ppmv (1-hour average),
respectively.18
The amount of process data included in the emissions test report is
fairly limited. Steam load varied from 80 to 103 percent of capacity during
the test. The thermal de-NOx system was operational throughout the testing
period except for one test run when the NOX levels were measured without
ammonia injection. This control system demonstrated NOX reduction in excess
of 40 percent. The report contains no information on combustion air operating
levels.
The Commerce facility has all of the design components of good
combustion in place. It is judged that the 1700°F (927°C) superheater inlet
temperatures correspond to a temperature at the fully mixed height which meets
the good combustion recommendations. The operation of the facility is
maintained by a fully automatic control system. The unit generates
electricity, operating at full load whenever possible. All of the
verification measures are in place to monitor continuous performance and
ensure good combustion.
3.1.1.7 Qlmstead County. Minnesota
The last mass burn waterwall MWC included in this discussion is the
Olmstead County, MN facility, which began operating in March 1987. The
facility consists of two 100 tpd (91 Mg/day) Riley-Takuma MWCs. Although
CDD/CDF emissions data are not available for the facility at this time, the
plant does monitor CO continuously, and average reported values are 31-54 ppmv
at 7 percent 02 (averaging time not reported).19 The testing program at
Olmstead County was a characterization test to determine the operating and
emissions performance of one of the units prior to undertaking a gas co-firing
performance evaluation. Sampling was performed at multiple points in the
furnace for 0?. COg, CO, and NOX. A series of 21 tests were performed with
variations in load, combustion airflow, and air distribution. Furnace
3-11
-------
temperatures were measured by suction pyrometry during each test. Temperature
profiles were also measured at the economizer outlet location.
One of the main conclusions in the study was that the use of overfire
air contributes to a reduction in CO emissions and at the same increases NOX
levels. The highest NOX emissions (191 ppmv) and the lowest CO emissions (31-
54 ppmv) occurred when overfire air was operated at 33 percent of total air,
and the lowest NOX (70 ppmv) and highest CO (123 ppmv) occurred when overfire
air was turned off.
The majority of good combustion elements are in place at Olmstead
County. However, there are only three plenums along the grate length, which
is typical for smaller capacity systems. The auxiliary fuel capacity is lower
than specified by the guidelines (10 percent versus 60 percent of full load).
but this will not affect baseline operating conditions. The low economizer
exit gas temperatures [typically 425°F (218°C)] should minimize the potential
for formation of CDD/CDF in the ESP.
3.1.2 Model Plant Baseline Emissions
The data used to establish baseline CDD/CDF emissions for mass burn
waterwall model plants are plotted in Figure 3-1. The data include values for
each individual sampling run. Facility averages are also plotted when
multiple runs are available. Five of the seven data sets include uncontrolled
emissions measured upstream of flue gas cleaning equipment. Average emissions
range from 27 ng/dscm (one run) at Commerce to 170 ng/dscm (5 runs) at
Millbury. Two of the data sets consist of controlled emissions measured in
the stack downstream of an ESP. Average emissions reported for these two
plants (Tulsa and Alexandria) are 36 ng/dscm and 53 ng/dscm. respectively.
The ESP operating temperature was not specified for Tulsa during the test
runs. The stack temperature at Alexandria was reported to be 342°F (172°C)
during testing, so it is assumed that the ESP operated at or below 400°F
(204°C), and that some amount of CDD/CDF removal occurred in the ESP.
The available uncontrolled data indicate that three of the four plants
tested achieved CDD/CDF emissions below 100 ng/dscm. However, two of the data
sets consist of a single sampling run. and these single runs may not
accurately represent an average CDD/CDF emission level. This is verified by
3-12
-------
s,
c
cr
15
D *
4»«Q
c
o
O
CL
CO
LU
90J8LULUOO
(Z860
(9861.)
o
o
c
c
o
"Jo
C
0)
"3
v>
CO
m
"5
c
3
CD
v>
w
CO
CO
a>
3
O)
o
in
CNJ
O
in
o o
m
(2O %L IB iuosp/6u)
3-13
-------
the multiple runs obtained at Marion County in 1987. Based on the available
data in Figure 3-1. it is assumed that all seven of the facilities are able to
achieve an uncontrolled baseline emission level of 200 ng/dscm COD/CDF. There
are insufficient data available to establish different baseline emission
levels for each of the mass burn waterwall model plants based on manufacturer
or size. Therefore. 200 ng/dscm was selected as a baseline emission level for
all three mass burn waterwall model plants.
The available data on CO emissions and uncontrolled PM emissions are
relatively consistent for the mass burn waterwall population. The CO data
reported for the seven plants included in this analysis varied from 4 ppmv (3-
hour average) at Pinellas County to 43 ppmv (unspecified averaging time) at
Olmstead County. Tulsa and Commerce report CO levels less than 30 ppmv with a
1-hour averaging time. Alexandria reports 18 ppmv on a 3-hour averaging time.
and Marion County reports 18 ppmv on a 4-hour averaging time. All of the
facilities tested can achieve 50 ppmv on a 4-hour average, which was selected
as a baseline value. Uncontrolled particulate emissions will vary according
to boiler design, air distribution, and waste characteristics. For example,
facilities that operate with high underfire/overfire air ratios or relatively
high excess air levels may entrain greater quantities of PM and have higher
uncontrolled emissions. Boilers with multiple passes that change the
direction of flue gas flow in the convective section may remove greater
quantities of entrained PM prior to entering flue gas cleaning equipment.
Lastly, the physical properties of waste being fed to a unit may impact the
amount of PM that becomes entrained. Despite these factors, a nominal
uncontrolled PM emission rate from a mass burn waterwall MWC is 2 gr/dscf
(4600 mg/dscm). The available data for the units discussed above ranges from
0.89 gr/dscf (2050 mg/dscm) at Marion County to 1.56 gr/dscf (3660 mg/dscm) at
Commerce. It is anticipated that all new mass burn waterwall MWCs will
achieve a baseline uncontrolled PM emission rate of 2 gr/dscf (4600 mg/dscm).
3.2 Split Flow Refractory MWCs
The mass burn refractory model plant is assumed to use the Volund
combustion technology. The Volund combustor is characterized by a split
furnace flow design which may be configured in two different ways (see Figures
3-2 and 3-3). The Volund systems are designed either with a refractory arch
which splits the flow, or with a flue gas overpass and a rotary kiln. The
kiln arrangement is generally used for medium and large size units [greater
3-14
-------
Figure 3-2. Volund Split Flow Furnace
Figure 3-3. Volund Split Flow Furnace with Rotary Kiln
3-15
-------
than 150 tpd (136 Mg/day)] while the refractory arch is used for small units.
If locally available waste is high in moisture, difficulty achieving good
burnout is possible and the rotary kiln is preferred. The kiln exposes the
waste to increased residence times at high temperatures and oxygen content.
resulting in improved burnout.
3.2.1 Emissions Data for Existing Facilities
3.2.1.1 McKav Bav. Florida
McKay Bay, FL. is one operating facility in the U.S. which uses the new
Volund technology. The facility consists of four 250 tpd (227 Mg/day) units
which use the rotary kiln configuration. The plant commenced operation in
1985, and compliance testing was performed on the four units in 1986.
Uncontrolled particulate emissions averaged 1.86 gr/dscf (4350 mg/dscm) and CO
emissions averaged 32 ppmv (individual units achieved 30, 35. 32. and 32
ppmv).14 Emissions of CDD/CDF were not reported for this plant. Very limited
process data are available for the test.
Table 3-2 summarizes recommended good combustion practices for
refractory wall MWCs using the split flow design. One specific requirement
that differs between refractory wall MWCs and waterwall MWCs is the excess air
operating range. Refractory wall MWCs typically operate at higher excess air
levels than waterwall MWCs due to the need to provide furnace cooling. An
evaluation of available information on the McKay Bay, FL. facility indicates
that several of the good combustion practice design criteria are not met,
including requirements for auxiliary fuel use and for exit gas temperatures.
The ESPs at McKay Bay typically operate near 550°F (288°C). so potential
exists for CDD/CDF formation to occur in the control devices. In addition.
the plant does not monitor CO. The combustion control scheme is based on
steam flows, temperatures, and air flows. The secondary air level controls
the temperatures exiting the furnace. Although combustor operating
temperatures are not specified, it is assumed that the units meet the
temperature requirements for good combustion. The McKay Bay units typically
operate at 60-135 percent excess air.
3-16
-------
TABLE 3-2. SPLIT FLOW REFRACTORY MWCS - PERFORMANCE ASSESSMENT
FACILITY
NUMBER OF UNITS - FGC
UNIT SIZE, tpd (Mg/day)
UNCONTROLLED EMISSIONS
CDD/CDF (ng/dscm)
CO (ppmv)
PM (mg/dscm)
CONTROLLED EMISSIONS
CDD/CDF (ng/dscm)
CO (ppmv)
COMBUSTION PARAMETERS
GOOD COMBUSTION
PRACTICE RECOMMENDATIONS
McKay Bay. FL
4 - ESP
250 (227)
Not Available (NA)
32
4280
NA
FACILITY DESIGN
AND OPERATING CONDITIONS
DESIGN
Temperature at fully
mixed height
Underfire air
Overfire air capacity
(not an operating
requirement)
Overfire air injector
design
Auxiliary fuel capacity
Exit gas temperature
OPERATION
Excess air
Turndown
Overfire air
Start-up procedures
Auxiliary fuel use
VERIFICATION
02 levels
CO
Temperature
Air distribution
Exit gas temperature
1800°F (982°C) mean
At least 4 plenums along
grate length (not
including kiln)
40% total air
Complete penetration/
coverage
As required to achieve
temperature limits
during start-up
<450°F (232°C)
6-14% 02 (dry)
80-110% design load
Penetration and coverage
of furnace cross section
Auxiliary fuel to
1800°F (982°C)
High CO. low temp;
start-up/shutdown
Monitor
Monitor (<100 ppmv at /% ()?))
Monitor
Monitor
Monitor
3-17
NA
3 plenums
Sidewall air
Quantity not specified
None
550°F (288°C)
8-12% 02
Unknown
Unknown
No auxiliary fuel
No auxiliary fuel
No
No
Yes
NA
NA
-------
3.2.1.2 Nvkoping. Denmark
A second set of emissions testing data is available from a Volund plant
located in Nykoping. Denmark.20 This facility includes two units with a
combined capacity of 168 tpd (153 Mg/day). The units, which began operating
in 1983, burn a mixture of domestic and non-hazardous industrial waste. Each
unit uses the refractory arch split flow configuration rather than a
refractory kiln. The majority of the combustion gases pass above the furnace
arch and are mixed with gases from the burnout grate. The units are equipped
with waste heat boilers. ESPs. and automatic combustion controls which
regulate furnace temperatures and steam production rates.
The emissions testing performed at Nykoping included CDD/CDF stack
sampling during start-up, normal operation, and variable boiler loads. Flue
gas oxygen concentrations were maintained between 7.4 and 9.7 percent, and
average furnace temperatures were approximately 940°C (1724°F). Temperatures
were measured above the transverse arch where the flow streams converge and
mix. Although the data are not tabulated for all runs, a typical CDD/CDF
emission level was reported to be 246.5 ng/Nm3 (sum of all CDD/CDF corrected
for spike recoveries).20 Graphical presentations of CO emissions are included
in the test report, but no values are tabulated for specific sampling runs.
It appears from the graphs that the units operate at CO levels between 50 and
200 ppmv. The authors of the test report concluded from the emission results
that mixing in the furnace has an important impact on the destruction of
CDD/CDF. and that concentrations of CO in the flue gases at well designed
plants can be used as a parameter for monitoring and control. There is
insufficient information available to assess the design and operation of the
Nykoping unit with the good combustion practice criteria, but the furnace
temperatures and oxygen levels reported are indicative of good combustion
conditions. The ESP operating temperature was not reported at Nykoping.
3.2.2 Model Plant Baseline Emissions
Based on the limited data available from the McKay Bay and Nykoping
MWCs. a baseline emission level of 300 ng/dscm CDD/CDF was established for the
model plant. Although all four of the combustors at McKay Bay achieved CO
emissions lower than 50 ppmv, a graphical presentation of CO data from
Nykoping indicates slightly higher levels during optimum combustion
3-18
-------
conditions. Therefore, baseline CO emissions were established at 100 ppmv.
The limited data available from the McKay Bay units supports an uncontrolled
particulate emission level of 2 gr/dscf (4600 mg/dscm).
3.3 Rotary Waterwall MWCs
The O'Connor Combustor Corporation was purchased by Westinghouse in
1986. The Westinghouse/O'Connor system design is unique; it includes a rotary
waterwall cylinder consisting of alternating watertubes and perforated steel
webs (i.e. a cylindrical, perforated membrane waterwall). Waste is fed to the
combustor by a dual ram feeder, and preheated combustion air is delivered
through the perforated webs by way of six windboxes located beneath the
barrel. The rotary section terminates within a waterwall boiler where bottom
ash is discharged and combustion gases pass through the boiler to produce
steam.
The first O'Connor MWC in the U.S. commenced operation in 1980 in
Gallatin, TN.21 Since Westinghouse purchased the O'Connor system, they have
started up two new facilities (Bay County. FL, and Dutchess County, NY).
Several more plants are in planning, permitting, or construction stages.
Parametric testing results obtained from Bay County and Gallatin have influ-
enced some aspects of Westinghouse's approach to designing new facilities.
For example, the Gallatin facility originally included rows of overfire
(tertiary) air ports in the radiant section of the boiler above the rear of
the rotary portion. In addition, an afterburning grate was located at the
discharge of the rotary section. Following field testing in 1984, the testing
engineers recommended that combustion air be eliminated from these locations
in the system. These modifications were made, but the air supplies were
reinstated after Westinghouse purchased O'Connor. Particulate, metals, and CO
emissions data are available from Gallatin, but they are not included in this
analysis because the facility is not representative of new
Westinghouse/O'Connor technology.
At Bay County the rotary portion of the combustor was designed with a
tapered end that protrudes several feet into the radiant section of the
boiler. The new Westinghouse/O'Connor plants are designed with a constant
diameter barrel; the end of the barrel is flush with the waterwalls rather
than extending into the radiant section of the boiler. In addition, the
3-19
-------
afterburning grate is a water cooled Detroit Stoker vibrating grate rather
than a stationary design.
3.3.1 Emissions Data for Existing Facilities
Initial continuous CO measurements made at Bay County indicated average
emissions of 100-200 ppmv. Westinghouse concluded that these relatively high
CO levels were due to (1) a low temperature zone below the protruding barrel
where mixing was not completed and (2) smoldering of ash on the afterburning
grate.22 Modifications were made to the combustion air distribution system,
including the addition of rows of tertiary air nozzles in the front and rear
boiler walls above the rotary combustor (similar to those originally in place
at Gallatin) and a small percentage of total air supplied beneath the after-
burning grate. These modifications reduced CO emissions to below 100 ppmv.
Results of continuous CO monitoring data from Bay County are reported to be 57
ppmv (3-hour average). 64 ppmv (7-hour average), and 83 ppmv (9-hour average)
during three separate testing periods.23 Westinghouse has gathered
uncontrolled and stack CDD/CDF data at Bay County, but the results have not
yet been reported.
The available Bay County emissions data and an assessment of combustor
performance are presented in Table 3-3. The good combustion practices for
rotary waterwall combustors require technology-specific application. For
example, the requirement for four separately adjustable air plenums includes
the 3x2 arrangement of windboxes in the rotary section of the unit and a
source of combustion air under the afterburning grate below the combustor
discharge. In addition, the O'Connor design typically operates at lower
excess air levels (40-80 percent) than conventional mass burners (40-130
percent). However, the basic approach to combustion control of emissions is
the same: good mixing is required at sufficient temperatures to destroy
organic emissions, and the potential for downstream formation of CDD/CDF must
be minimized.
3.3.2 Model Plant Baseline Emissions
The Bay County. FL CDD/CDF emission results have not yet been reported;
thus, no measured data are currently available to support baseline emission
estimates from Westinghouse/O'Connor combustors. Therefore, baseline
uncontrolled CDD/CDF emissions of 300 ng/dscm were established for the lll(b)
3-20
-------
TABLE 3-3. MASS BURN ROTARY WATERWALL MWCS - PERFORMANCE ASSESSMENT
FACILITY
NUMBER OF UNITS - FGC
UNIT SIZE, tpd (Mg/day)
UNCONTROLLED EMISSIONS
CDD/CDF (ng/dscm)
CO (ppmv)
PM (mg/dscm)
CONTROLLED EMISSIONS
CDD/COF (ng/dscm)
CO (ppmv)
COMBUSTION PARAMETERS
GOOD COMBUSTION
PRACTICE RECOMMENDATIONS
Bay County. FL
2 - ESP
255 (232)
Not Available (NA)
60-80
NA
NA
60-80
FACILITY DESIGN
AND OPERATING CONDITIONS
Temperature at fully
mixed height
Underfire air
Overfire air capacity
Tertiary air design
Auxiliary fuel capacity
Exit gas temperature
OPERATION
Excess air
Turndown
Tertiary air
Start-up procedures
Auxiliary fuel use
VERIFICATION
02 levels
CO
temperature
Air distribution
Exit gas temperature
1800°F (982°C) average
4 plenums, including
one at burnout grate
Sum of secondary and
tertiary air designed to
supply 40% of total air
Complete coverage and
penetration
That required to achieve
temperature limits
during start-up
<450°F (232°C)
3-9% 02 in flue gas (dry)
80-110% design load
Complete coverage and
penetration
Auxiliary fuel to
design temperature
High CO. low temp;
start-up/shutdown
Monitor
Monitor «100 ppmv at 7% 02)
Monitor
Monitor
Monitor
3-21
1400°F (760°C) at inlet
to convective section
4 plenums (one at
afterburning grate)
Confidential
Confidential
Oil - 40% load
450°F (232°C)
5-57% 02 (wet)
30% minimum
Not achieved
Use steam preheat from
adjacent combustor
NA
Yes
Yes
Yes
UF. OF. tertiary
Yes
-------
model plant based on engineering judgement. Based on the results reported
from Bay County, baseline CO emissions were established at 100 ppmv.
Uncontrolled PM emissions are assumed to be 2 gr/dscf (4600 mg/dscm).
consistent with uncontrolled PM measurements at conventional mass burn
waterwall facilities.
3.4 Modular Excess Air MWr.s
The population of planned and projected modular excess air MWCs consists
of several distinctly different designs. The three design types identified
with facilities in planning, permitting, or construction stages include:
• Vicon/Enercon (7 known projects)
• Cadoux (3)
• Basic (1)
All three of these designs utilize multiple combustion chambers; Vicon/Enercon
and Basic also employ flue gas recirculation (FGR). One key difference in the
physical designs is that the Basic primary chamber is a membrane waterwall,
while Vicon/Enercon and Cadoux use refractory wall combustors with separate
waste heat boilers.
3.4.1 Existing Facilities Emissions Data
3.4.1.1 Pittsfield. Massachusetts
There are three sets of CDD/CDF data available from modular excess air
MWCs. The first data set includes the parametric test results obtained from a
research program conducted at the Vicon/Enercon facility in Pittsfield. MA.6
A facility equipment schematic is shown in Figure 3-4. The plant comprises
three 120 tpd (109 Mg/day) units and two waste heat boilers. Testing was
performed with two of the three units in operation, which is the normal
operating condition for the facility. Each boiler exhausts to an electrified
granular bed (EGB) filter for removal of PM. The EGBs typically operate at
475°F (246°C). Organic emissions, including CDD/CDF, PCBs, chlorophenols. and
cnlorobenzenes, were measured over a large range of operating conditions and
while firing various fuels (MSW. PVC spiked MSW, PVC free waste).
3-22
-------
t-
<
V)
V)
o
X
T3
O
z:
oo
o>
3-23
-------
Table 3-4 presents a summary of average CDO/CDF and CO emissions
reported for each operating condition investigated in the parametric test.
The temperature specified for each test condition is measured at the exit of
the secondary combustion chamber. Each test condition consisted of two
individual sampling runs, with the exception of the 1400°F condition, which
included only one run. Sampling for CDD/CDF was performed at the boiler
outlet for each operating condition. CDD/CDF stack sampling was performed
during two conditions (1800°F-MSW and 1800°F-MSW + PVC). and CDD/CDF samples
were gathered at the tertiary duct (just upstream of the boiler) during the
eight other conditions. Carbon monoxide emission levels were measured in the
tertiary duct and at the boiler outlet during all runs. Fairly extensive
process monitoring was performed during this test program, including
measurement of airflows and temperatures at various locations in the system.
Testing was performed for the suspected precursors of CDD/CDF (PCBs,
chlorobenzenes, chlorophenols). Continuous monitors were maintained to
measure Og, CC"2, CO. SOg, NOX, HC1. and THC at various locations in the
system. Control of combustion temperatures was maintained by modulation of
feed rates and recirculated flue gas and. to a lesser extent, fresh airflows.
One of the main conclusions made in the data analysis was that CDD/CDF
emission levels were not generally affected by the different waste character-
istics evaluated in the program. However, as expected, emissions of HC1 were
noticeably affected by PVC content in the waste. In most cases. CDD/CDF
concentrations increased at each sampling location as the flue gases passed
through the system. The temperatures at the tertiary duct sampling location
are nearly equivalent to the target values specified in each sampling
condition. The sampling point was upstream of the flue gas recirculation
injection point (see Figure 3-4). The amount of recirculated (tempering) air
injected into the tertiary duct controls the boiler inlet gas temperature,
which varied from approximately 1100°F (593°C) during the 1300°F-MSW condition
to approximately 1400°F (760°C) during normal operating conditions (1800°F-
MSW). The average boiler outlet gas temperature varied from 460 to 540°F (255
to 282°C). Therefore, the flue gases pass through the critical CDD/CDF
formation temperature (approximately 300°C) in the boiler. This is reflected
by the increases in CDD/CDF concentrations between the tertiary and boiler
outlet sampling locations. This result was observed during six of the eight
conditions. The formation rate appears to be higher than actually may be
occurring, because the gas stream at the boiler outlet location contains gases
recirculated from ahead of the APCD. Increased concentrations of COD were
3-24
-------
TABLE 3-4. PITTSFIELD, MA MODULAR EXCESS AIR MWC
EMISSIONS TEST DATA
TESTING CONDITION
1300°F - MSW
1400°F - MSW
1550°F - MSW
1550°F - MSW + H20
1800°F - MSW
1800°F - MSW. low 02
1800°F - MSW + PVC
1800°F - PVC free
1800°F - PVC free + PVC
1800°F - PVC free + H20
TERTIARY DUCT
CDD/CDF
(ng/dscm)
112
18
15
57
-
76
-
31
14
48
CO
(ppmv)
201
44
9
17
4
7
6
1
6
8
BOILER OUTLET
CDD/CDF
(ng/dscm)
403
40
57
21
94
165
148
71
87
28
CO
(ppmv)
148
22
15
14
12
9
7
9
13
7
STACK
CDD/CDF
(ng/dscm)
-
-
-
-
154
-
261
-
-
-
3-25
-------
measured between the two sampling points during all but two conditions (1550°F-
MSW + H20. and 1800°F-PVC free + H20). The influence of water on the CDD/CDF
formation mechanism must be investigated further in order to draw conclusions
related to this observation at Pittsfield. No PM sampling was performed
during any of the test runs. However, waste moisture content may have reduced
the amount of PM that was entrained in the fly ash and available for
downstream catalytic reactions to occur. A normal sootblowing cycle was
reportedly performed for the boiler during each 4-hour sampling run. The
facility was reportedly scheduled for an annual maintenance shutdown 2 weeks
after the completion of testing; the condition of the plant was considered
normal during testing (no special maintenance was performed prior to
initiating the program).
Paired runs gathered during normal operating conditions (1800°F-MSW)
provided an average CDD/CDF emission rate of 94 ng/dscm at the boiler outlet
(154 ng/dscm at the stack). At the 1500°F-MSW test condition, the total
CDD/CDF emissions averaged 57 ng/dscm at the boiler outlet (no stack measure-
ments available). As secondary chamber temperatures were decreased to 1300°F.
the average CDD/CDF emission rate increased to 403 ng/dscm at the boiler
outlet. The low temperature runs were performed for experimental purposes and
are not expected to be encountered during normal operating conditions.
Stack testing was performed at Pittsfield during two operating
conditions (1800°F-MSW and 1800°F-MSW + PVC). Concentrations of CDD/CDF
increased by 64 and 76 percent, respectively, from the boiler outlet location
to the stack. Average boiler outlet gas temperatures ranged from 472 to 536°F
(250 to 280°C) during these four sampling runs.
During all the aforementioned testing runs, CO emissions were measured,
and average emission levels at the boiler outlet did not exceed 22 ppmv (4-
hour average) except when operating at 1300°F, when 148 ppmv was measured.
The extensive emissions and process data generated at Pittsfield
demonstrate that sufficiently high temperatures and adequate mixing conditions
are present to minimize CDD/CDF and CO emissions at normal operating
conditions. The low emission levels measured at Pittsfield confirm that the
units have good combustion practices in place.
3-26
-------
3.4.1.2 Piaeon Point. Delaware
A second data set available for the Vicon/Enercon facility is Pigeon
Point. DE. This plant comprises five 120 tpd (109 Mg/day) units that fire a
mixture of MSW and RDF, with ESP controls. The compliance test at Pigeon
Point was conducted in two phases. Phase I consisted of HC1. $03, CO. and PM
measurements made in the stack. Particulate testing was also performed at the
ESP inlet location for all four units. Phase II involved stack sampling for
CDD/CDF and heavy metals (Pb. Hg. Be. Ni, As. Cd. Cr).
The three-run average uncontrolled PM levels for each of the four flues
tested were 1.03 gr/dscf (2370 mg/dscm), 1.03 gr/dscf (2370 mg/dscm). 0.90
gr/dscf (2070 mg/dscm), and 0.43 gr/dscf (990 mg/dscm).24 One of the flues
discharges gases from two combustors. Average flue gas temperatures entering
the ESP ranged from 393 to 433°F (200 to 223°C). Due to the low particulate
concentrations and the low operating temperature, it is doubtful that
substantial CDD/CDF formation would occur in the ESP.
Three CDD/CDF sampling runs were performed in the stack of unit #2. The
average emissions were reported to be 105 ng/dscm.24 The average stack
temperature was reported to be 374°F (190°C). The CO data included in this
test were measured by ORSAT analysis and were reported to be 0.0 percent by
volume. The CDD/CDF concentrations in the ESP fly ash are also reported for
each sampling run. Very limited process data are available to use in
characterizing the operation of the Pigeon Point facility during testing. The
plant attempts to feed a mixture of 5 pounds RDF per pound of MSW. and it
appears that this ratio was maintained during the tests. Based on the
assumption that the combustor design and operation are similar to that at
Pittsfield. it can be concluded that good combustion practices are in place at
Pigeon Point. The measured emission levels from Pigeon Point and Pittsfield
confirm the good performance of the Vicon/Enercon design. The consistency of
the CDD/CDF data with that measured at Pittsfield also indicate that CDD/CDF
emission levels are more dependent on combustion technology than on waste feed
characteristics.
3.4.1.3 Alexandria. Minnesota
A third data set is available from a facility using the Cadoux design.
Emissions testing was performed at the Pope/Douglas Waste-to-Energy Facility
3-27
-------
in Alexandria. MN. in July 1987. This plant began operating in May 1987 using
two 38 tpd (35 Mg/day) Cadoux modular excess air combustors. Both units are
equipped with an ESP. Average CDD/CDF emissions at the stack were reported to
be 446 ng/dscm.25 The continuous monitoring results indicated that average CO
emissions were 24 ppmv (1-hour average). The average flue gas temperature at
the ESP inlet sampling location ranged from 490 to 503°F (254 to 262°C).
These values are in the temperature window where CDD/CDF formation has been
observed. Therefore, it is assumed that CDD/CDF at the ESP inlet was lower
than the concentrations in the stack. The CDD/CDF and CO measurements were
made on Unit #2. Particulate sampling was also performed in the stack of both
units.
Thermocouples were installed in Unit #1 at two locations in the primary
combustion chamber. The first was located on the side wall at the grate
level, and the second was in the top of the chamber. The average temperature
at the grate varied from 1300 to 1650°F (704 to 899°C) and the temperature in
the upper furnace ranged from 1770 to 1990°F (965 to 1088°C). Charge rates
were measured during the testing program and both units were operating between
100 and 145 percent rated capacity. Oxygen levels were also measured in an
eight-point traverse at the combustor outlet, and average concentrations were
14.8 percent at Unit #1 and 12.8 percent at Unit #2. These values equate to
approximately 245 percent and 160 percent excess air. respectively.
3.4.2 Model Plant Baseline Emissions
Baseline emissions for the model plant are based largely on the
performance of the two Vicon/Enercon MWCs at Pittsfield and Pigeon Point. The
emissions data from these systems and the Cadoux facility at Pope/Douglas
County are plotted together in Figure 3-5. The only data included from the
Pittsfield parametric test are those runs which took place under normal
operation, firing 100 percent MSW. The off spec, low temperature runs are not
expected to be encountered in normal operation. Baseline emissions are
assumed to be 200 ng/dscm CDD/CDF and 50 ppmv CO. Uncontrolled PM emissions
are assumed to be 2 gr/dscf (4600 mg/dscm), which is an average value for mass
burn systems.
3-28
-------
cc
ll
I
0>
w
IB
m
-'Id
Q_
c
o
"•53
(0
Q>
i
Q)
"3J
w
(Q
CO
U)
X
ill
k.
CO
CO
888888
to in f co CM i-
(zo %L IB uiosp/6u)
3-29
-------
3.5 Mass Burn Modular Starved Air MWCs
There are at least four manufacturers of starved air modular MWCs with
units in planning, permitting, and/or construction stages (Consumat. Clean
Air. Joy. and John Zink). Although modular starved air facilities comprise
the largest category of existing MWCs in total facility numbers, there are
very few identified in the planned and projected population. This may be due
to the short amount of time required to construct a modular facility; the
current list of planned and projected plants may represent shorter term
projections for this population.
Generic descriptions of modular starved air MWCs are included in EPA's
MWC Report to Congress. Refinements in combustor design and controls continue
to be made in the new starved air MWC population. Consumat Systems has
increased the volume of the secondary combustion chamber in their more recent
designs.26 This modification is intended to increase residence time of gases
for better mixing. A second change in the Consumat design is related to
control of the charging rate. Originally, the charging rate was based on a
pre-established interval which was adjusted manually. The loader operator was
given a signal every 5-10 minutes to add another charge to the loader. The
size of the charge (heavy, normal, or light) was also indicated by a signal
which was based on the primary chamber temperature. The operator adjusted the
size of individual charges based on the signal, increasing the size if the
temperature was below the setpoint and decreasing the charge if the setpoint
temperature was exceeded. In the newer Consumat designs, the time between
charges is automatically adjusted based on a feedback signal from the
temperature controller. The operator does not have to adjust the size of each
charge manually: instead the controller will automatically lengthen or shorten
the time between charges, eliminating dependency on operator judgment. This
feature may contribute to lower organic emissions, because it is hypothesized
that CDD/CDF levels correlate with the number of CO spikes. The adjustable
feed control sequence may provide less frequent spikes in CO emissions and
lower uncontrolled trace organic emissions.
The Joy design is similar to Consumat in that transfer rams are used to
move waste through the primary combustion chamber. Some unique features in
the Joy system are an underflre air and steam injection system for temperature
control, and a modulating temperature controller which automatically adjusts
3-30
-------
the induced draft (ID) fan damper. The air/steam injection system in the
primary chamber is designed to reduce high localized hearth temperatures to
minimize clinker formation and minimize vaporization of inorganics (metals).
In addition, alternating the exposure of the waste to underfire air and steam
reportedly maximizes the conversion of fixed carbon content of the waste to
C0£. The automatic control of the ID fan damper results in fewer furnace
upsets due to pressure imbalances, changes in draft, or changes in total
combustion gas flow. When these upsets occur and temperatures increase, a
thermocouple in the dumpstack senses the increase in temperature and the
controller drives the ID fan damper to a more open position, resulting in
improved system stability. If the thermocouple senses an immediate decrease
in temperature the ID fan damper is driven to a more closed position, which
reduces air flows and boosts operating temperatures.
3.5.1 Existing Facilities Emissions Data
CDD/CDF emissions data are available from several modular starved air
MWCs. The most recent data has been gathered by Environment Canada at Prince
Edward Island (PEI) and by New York DEC at the Oswego, Oneida, and Cattaraugus
County MWCs. Uncontrolled data are available from PEI, Cattaraugus County.
and Oswego. The available emissions data are included in Table 3-5 along with
a comparison of individual plant performance relative to the good combustion
practice recommendations for modular starved air MWCs.
3.5.1.1 Prince Edward Island
Emissions testing was performed at PEI at four operating conditions
(normal, long feed cycle, high secondary chamber temperature, and low
secondary chamber temperature). The facility consists of three Consumat CS-
1600 combustors, each rated at 36 tpd (33 Mg/day). The combustors exhaust to
a common waste heat recovery boiler and then to a stack without further
emissions control. A process schematic of the facility is provided in Figure
3-6. Sampling was performed at the boiler inlet location and in the stack.
The primary operating variables were primary and secondary combustion
temperatures and feed cycle. Three sampling runs were performed for each
condition. The average CDD/CDF and CO data are presented in Table 3-6 for
each operating condition.&
3-31
-------
TABLE 3-5. MODULAR STARVED AIR MWCS - PERFORMANCE ASSESSMENT
PAGE 1 OF 2
FACILITY
NUMBER OF UNITS - FGC
UNIT SIZE, tpd (Mg/day)
UNCONTROLLED EMISSIONS
CDD/CDF (ng/dscm)
CO (ppmv)
PM (mg/dscm)
CONTROLLED EMISSIONS
CDD/CDF (ng/dscm)
CO (ppmv)
COMBUSTION PARAMETERS
Prince Edward Island
3 - None
36 (33)
409
62
225
GOOD COMBUSTION
PRACTICE RECOMMENDATIONS
FACILITY DESIGN
AND OPERATING
CONDITIONS
Temperature at fully
mixed height
Secondary air capacity
(not an operating
requirement)
Secondary air injector
design
Auxiliary fuel capacity
Exit gas temperature
OPERATION
Excess air
Turndown
Secondary air
Start-up procedures
Auxiliary fuel use
VERIFICATION
02 levels
CO
Temperature
Air distribution
Exit gas temperature
1800°F (982°C) average
80% total air
As required to achieve
temperature limits
during start-up
<450°F (232°C) at PM
control device inlet
6-12% 02 (dry)
80-110% design load
80% total air
On auxiliary fuel to
design temperature
High CO. low temp;
start-up/shutdown
Monitor
Monitor (<50 ppm at 7% 02)
Monitor
Monitor
Monitor
1832°F (1000°C)
(secondary chamber)
NA
That required for penetration NA
and coverage
NA
363°F (184°C)
12% 02
NA
NA
NA
NA
No
No
Primary and
secondary chamber
No
No
3-32
-------
TABLE 3-5. MODULAR STARVED AIR MWCS - PERFORMANCE ASSESSMENT
PAGE 2 OF 2
FACILITY
NUMBER OF UNITS - FGC
UNIT SIZE, tpd (Mg/day)
UNCONTROLLED EMISSIONS
CDD/CDF (ng/dscm)
CO (ppmv)
PM (mg/dscm)
CONTROLLED EMISSIONS
CDD/CDF (ng/dscm)
CO (ppmv)
COMBUSTION PARAMETERS
Cattaraugus County, NY
3 - None
38 (35)
345
NA
Onelda County, NY
4 - ESP
50 (45)
FACILITY DESIGN
AND OPERATING CONDITIONS
462
FACILITY DESIGN
AND OPERATING CONDITIONS
DESIGN
Temperature at fully
mixed height
Secondary air capacity
(not an operating
requirement)
Secondary air injector
design
Auxiliary fuel capacity
Exit gas temperature
OPERATION
Excess air
Turndown
Secondary air
Start-up procedures
Auxiliary fuel use
1800-2000°F (983-1093°C)
(secondary exit)
At least 40% total air
NA
Gas - 30% load
502°F (216°C) (stack)
NA
NA
40% of total air
On gas to 1800°F (983°C)
in secondary
Start-up
1800°F (983°C)
(secondary exit)
NA
NA
Gas - 100% (not used)
400-450°F (204-232°C)
(boiler outlet)
NA
NA
NA
Not used
None
VERIFICATION
02 levels
CO
Temperature
Air distribution
Exit gas temperature
No
No
Primary and secondary
chamber
Primary air
Yes
No
No
Primary and secondary
chamber
No
Yes
3-33
-------
"'"^y.r^sa
•a
c
3-34
-------
TABLE 3-6. PERFORMANCE TEST DATA
PRINCE EDWARD ISLAND MWC
CONDITION
Normal
Long Cycle
High Secondary
Low Secondary
BOILER INLET
CDD/CDF
(ng/dscm)
NA
0
1
42
TEMPERATURE
op oC
1544 840
1544 840
1922 1050
1364 740
STACK
CDD/CDF
(ng/dscm)
409
441
198
424
CO
(ppmv)
62
39
38
53
TEMPERATURE
op oC
363 184
363 184
361 183
266 130
3-35
-------
The data in Table 3-6 indicate that COD/CDF emission levels in the stack
are partially due to formation that occurs in the lower temperature portions
of the system. With the exception of the low secondary temperature
conditions. CDD/CDF emissions are near zero at the boiler inlet. The higher
emission levels during the low secondary temperature condition likely result
from insufficient temperatures to provide destruction of CDD/CDF and
precursors. Secondly, the operating variable that had the most noticeable
effect on CDD/CDF stack emissions was the high secondary chamber temperature.
These data provide support for the combustion temperature requirements in the
MWC recommendations.
3.5.1.2 Cattaraugus County. New York
The second set of CDD/CDF emissions data was gathered at the Clear Air
facility in Cuba (Cattaraugus County), NY. The plant consists of three 38 tpd
(35 Mg/day) units that began operating in 1983. The plant has heat recovery
but uses no flue gas cleaning device. Two CDD/CDF sampling runs are available
from testing performed by New York State in 1984, and average emissions were
reported to be 345 ng/dscm.27 There are no CO data available with the test
results. Primary chamber temperatures were approximately 2200-2300°F (1204-
1260°C) and secondary chamber temperatures were maintained near 2000°F
(1093°C) during testing.
3.5.1.3 Oneida County. New York
Oneida County is another Clear Air facility which comprises four units
at 50 tpd (45 Mg/day) each. The plant has heat recovery in place and is
equipped with an ESP. Stack testing was performed by New York State DEC in
1985. Average CDD/CDF emissions at Unit #1 were 462 ng/dscm.15 The
temperature at the ESP inlet was 458°F (237°C). Primary chamber temperatures
were approximately 1600-1800°F (871-982°C) and secondary chamber temperatures
were 1700-2000°F (927-1093°C) during testing.
3.5.1.4 Additional Data
Uncontrolled PM emissions data are available from Tuscaloosa. AL [0.07
gr/dscf (160 mg/dscm)]^, PEI [0.098 gr/dscf (225 mg/dscm)]*. and Windham. CT
[0.07 gr/dscf (160 mg/dscm)].28 CO emissions are available from PEI (38-62
ppmv)5 and from Red Wing, MN (2 pprnv).™
3-36
-------
The available emissions data provide evidence that relatively low
CDD/CDF concentrations can be achieved by modular starved air MWCs. The key
conditions that lead to low emissions are the same as specified for other
technologies: achieve good mixing at adequate temperature and minimize the
conditions that lead to downstream formation of CDD/CDF. Starved air MWCs can
achieve adequate secondary chamber temperatures by control of air flows. The
fully mixed location in a modular starved air MWC can be defined at the exit
of the secondary combustion chamber. The available data also indicate that
total elimination of downstream formation of CDD/CDF may not be feasible.
However, systems should be designed and operated in a manner which minimizes
the potential for these occurrences.
3.5.2 Model Plant Baseline Emissions
Figure 3-7 presents the available CDD/CDF emissions data in graphical
form. Based on the available emissions data and on judgements that design
improvements made to new systems (as discussed above) will result in lower
emission levels, baseline uncontrolled CDD/CDF emissions were established at
300 ng/dscm. Based on available data from several existing facilities, newer
facilities are also expected to be able to achieve uncontrolled particulate
matter emissions of 0.1 gr/dscf (230 mg/dscm) and CO emissions of 50 ppmv.
3.6 Model Plant Performance Estimates - RDF Combustors
Refuse-derived-fuel (RDF) combustion technology includes the use of
conventional RDF spreader stokers, fluidized bed combustors (FBC). and RDF co-
firing in coal fired boilers. The RDF model plants in this analysis include
two spreader stokers and two FBC facilities. One of the spreader stoker model
plants burns 100 percent RDF and one burns a 50/50 mixture of RDF and wood
waste. Based on projections of the new population of FBCs. one circulating
fluidized bed (CFB) model plant and one bubbling fluidized bed model plant
were developed. Both of the fluidized bed models are assumed to burn 100
percent RDF.
3.6.1 RDF Spreader Stokers
The population of planned and projected RDF fired spreader stoker
boilers is anticipated to include two major manufacturers--Babcock & Wilcox
3-37
-------
c
(T
o "2
g» -g
? =6
< JE
•
•
0 0
0 0
h- to
a •
• D 4
• • Q
CD
"5
1
4» Q
D •
*
•
•
)
UJ TO
D. ~
.1 i
o •.'
Z r*"
£
3
O)
88888° ""
m •* eo CM r-
(zo
uiosp/6u)
3-38
-------
(B&W) and Combustion Engineering (CE). Both of these manufacturers have
developed new RDF combustors with unique design features.
Figure 3-8 shows the configuration of the new B&W design. The lower
portion of the boiler is pinched to provide arches in the front and rear wall
and a row of overfire air nozzles is located at each arch. Overfire air is
also injected below the arches. The configuration is designed to provide a
well mixed region in the lower furnace above the traveling grate. The first
facility to use this design is the Maine Energy Recovery Company (MERC) plant
at Biddeford, ME. Another design feature which is characteristic of new RDF
fired MWCs is a metered feeding system. A Detroit Stoker metered feeding
system (shown in Figure 3-9) is also used at MERC. The metered feeding system
consists of two RDF hoppers, a ram feeder, and a variable speed conveyor which
transports RDF to air swept distributors at a constant rate, thus providing a
consistent fuel feed rate and uniform distribution of waste on the traveling
grate.
A second new RDF boiler design being supplied by Combustion Engineering
includes multiple, separately controllable underfire air plenums (Figure 3-
10). This design provides the operator with the ability to vary underfire air
distribution to each region of the waste bed based on the distribution of fuel
on the grate. A second unique design feature in the CE boilers is the
overfire air system. There are tangential overfire air jets located in the
furnace corners, each of which includes three levels of separately
controllable nozzle banks. There are also three rows of overfire air nozzles
on the furnace walls (one row on the front wall and two rows on the rear
wall). Although the wall fired air is normally operated when burning coal.
the tangential and wall overfire air systems can be fired simultaneously.
Metered feeding and segmented underfire air supplies are both designed
to accomplish the same objective: to provide uniform stoichiometries in the
grate/fuel bed region of the furnace. This is a necessary requirement for
achieving good combustion. The objective can be accomplished either by
biasing the supply of combustion air to the portion of the grate where the
waste is concentrated, or by distributing the waste evenly on the grate with
underfire air delivered from a single plenum.
3-39
-------
Overfire Air
Undergrate Air
Figure 3-8. MERC - B&W Boiler Design
3-40
-------
E
O)
GO
X3
QJ
O)
^
O
4->
•r-
O
Ol
Q
(-
3
Dl
3-41
-------
RDF
Distributors
Tangential
Overflre Air
Crate Surface
Drive Shaft
Undergrate
Air Compartment
Sifting Screw
Conveyor
Figure 3-10. Mid-Connecticut CE RDF Fired MWC
3-42
-------
3.6.2 Existing Facilities Emissions Data
A summary of emissions test data is presented in Table 3-7 for RDF fired
facilities that are judged to be representative of the planned and projected
population.
3.6.2.1 Biddeford. Maine
The Biddeford. ME. MWC was tested for CDD/CDF by EPA in December 1987.
The plant comprises two boilers, each rated at 350 tpd (318 Mg/day) RDF.
Emissions control on each unit is achieved by a cyclone, a spray dryer, and a
baghouse. Emissions were measured at the spray dryer inlet (downstream of the
cyclone) in conjunction with compliance testing performed in the stack. The
unit that was tested was operating at full load during the test. Although an
RDF/wood mixture is fired during normal operating conditions. 100 percent RDF
was burned during the test. Three CDD/CDF sampling runs were performed and
average unabated emissions were 903 ng/dscm.29 Inlet particulate emissions
were 3.56 gr/dscf (8190 mg/dscm). and average CO emissions were 81 ppmv. The
boiler was operated at approximately 65-70 percent excess air during the test
with an overfi re/underfi re ratio of 56/44. Each test run was 4 hours'
duration. The average temperature at the spray dryer inlet location was 374°F
(190°C).
3.6.2.2 Red Wind. Minnesota
The Red Wing facility comprises two 35-year-old coal-fired spreader
stoker boilers that have been retrofitted to burn 100 percent RDF.30 The
boilers were enlarged by extending the furnace down into the basement.
removing the coal bottom ash hoppers, lowering the stokers, and adding a 14-
foot (4.3 m) waterwall extension fabricated from membrane panels with high
nickel alloy weld overlay. The existing tube and tile upper furnace was
connected to the new membrane walls by installation of a transition header. A
new multilevel, multipoint, heated overfire air system was installed to ensure
good mixing.
The facility was tested for CDD/CDF in 1987. Emissions at the ESP inlet
were reported to be 60 ng/dscm and average stack emissions were 28 ng/dscm.31
Inlet particulate emissions were reported to be 2.13 gr/dscf (4900 mg/dscm)
3-43
-------
TABLE 3-7. RDF FIRED SPREADER STOKERS - PERFORMANCE ASSESSMENT
PAGE 1 of 2
FACILITY
NUMBER OF UNITS - FGC
UNIT SIZE, tpd (Mg/day)
UNCONTROLLED EMISSIONS
CDD/CDF (ng/dscm)
CO (ppmv)
PM (mg/dscm)
CONTROLLED EMISSIONS
CDD/CDF (ng/dscm)
CO (ppmv)
COMBUSTION PARAMETERS
GOOD COMBUSTION
PRACTICE RECOMMENDATIONS
Red Wing, MN
2 - ESP
360 (318)
NA
127
4900
28
99
FACILITY DESIGN
AND OPERATING CONDITIONS
DESIGN
Temperature at fully
mixed height
Underfire air control
Overfire air capacity
Overfire air injector
design
Auxiliary fuel capacity
Exit gas temperature
OPERATION
Excess air
Turndown
Overfire air
Start-up procedures
Auxiliary fuel use
VERIFICATION
02 levels
CO
Temperature
Air distribution
Exit gas temperature
1800°F (982°C) mean
As required for uniform
bed stoichiometry
40% total air
Penetration and coverage
As required to achieve
temperature limits
during start-up
<450°F (232°C) at PM
control device inlet
3-9% 02 (dry)
80-110* of design
Coverage and
penetration
Auxiliary fuel
High CO. low temp;
start-up/shutdown
Monitor
Monitor (100 ppm at 7% 02)
Monitor
Monitor
Monitor
>1800°F (982°C) inlet
to first convective
section
2 plenums
50* total air
100% load
260-450°F (127-232°C)
7-11% 02
40% minimum
50% at full load
20% at minimum load
Gas
Start-up
Yes
Yes
Yes
Yes
Yes
3-44
-------
TABLE 3-7. RDF FIRED SPREADER STOKERS - PERFORMANCE ASSESSMENT
PAGE 2 of 2
FACILITY
NUMBER OF UNITS - FGC
UNIT SIZE, tpd (Mg/day)
UNCONTROLLED EMISSIONS
CDD/CDF (ng/dscm)
CO (ppmv)
PM (mg/dscm)
CONTROLLED EMISSIONS
CDD/CDF (ng/dscm)
CO (ppmv)
COMBUSTION PARAMETERS
Biddeford. ME
2 -
300 (272)
903
81
8190
FACILITY DESIGN
AND OPERATING CONDITIONS
DESIGN
Temperature at fully
mixed height
Underfire air control
Overfire air capacity
Overfire air injector
design
Auxiliary fuel capacity
Exit gas temperature
OPERATION
Excess air
Turndown
Overfire air
Start-up procedures
Auxiliary fuel use
VERIFICATION
02 levels
CO
Temperature
Air distribution
Exit gas temperature
Not Available (NA)
Metered fuel feeding
60% total air
Gas - 40% load
374°F (190°C) at PM control
device inlet
7% 02
40% minimum (on gas)
60% total air
Gas to 1800°F (982°C)
Routinely co-fire, start-up, shutdown
Yes
No
Not currently measured
NA
Yes
3-45
-------
and average CO levels were 127 ppmv (5-hour average). The Red Wing units are
designed to operate at 65 percent excess air with a 50/50 overfire/underfire
air ratio. The average excess air level was 62 percent during CDD/CDF
testing, and the air distribution was not specified. The design ESP inlet
temperature is 420°F (216°C) and the average operating value was 425°F
(218°C).
After the data were subjected to an EPA QA/QC review, the inlet CDD/CDF
data were invalidated. Therefore, only the data measured in the stack were
included in this analysis.
3.6.3 Model Plant Baseline Emissions
Based on the measured data from Biddeford, the baseline CDD/CDF emission
levels for new RDF units have been established at 1000 ng/dscm. Baseline CO
values of 100 ppmv are established for the model plants. This CO emissions
level is expected to be achievable by new RDF fired units in a short term (4-
hour) test. Average uncontrolled PM emissions of 4 gr/dscf (9200 mg/dscm)
were assumed for the model plants based on the Biddeford data.
3.7 RDF/Wood Co-Fired MWCs
At this time there are no data available to establish baseline CDD/CDF
emissions for RDF boilers that fire mixtures of RDF and wood. Although
untreated wood will probably have lower chlorine contents than RDF, the ring
structure organics emitted from the combustion of wood could possibly
contribute to CDD/CDF formation. Due to the absence of data to indicate
otherwise, the same baseline emission levels are assumed for spreader stokers
firing mixtures of wood and RDF: 1000 ng/dscm CDD/CDF. 100 ppmv CO, and 4
gr/dscf (9200 mg/dscm) PM.
3.8 Fluidized Bed Combustors
There are two FBC model plants; one bubbling bed facility and one
circulating fluidized bed (CFB) plant. The actual growth rate of CFB
combustion units in the U.S. is questionable at this time. Several
communities reportedly have plans for new CFB units, but very few are in
construction or permitting stages.32
3-46
-------
3.8.1 Existing Facilities Emissions Data
The two fluidized bed MWCs in the U.S. (Duluth. MN and Lacrosse, WI) are
both bubbling bed designs. Each plant has been tested for CDD/CDF.
Discussion of test results is provided below.
3.8.1.1 Duluth. Minnesota
The Duluth facility normally combusts a mixture of RDF and sewage sludge
[120 tpd (109 Mg/day) fluff RDF and 345 tpd (314 Mg/day) sludge at 18 percent
solids]. Emissions are controlled by a venturi scrubber. Testing was
performed at full operating load. Total tetra-octa CDD/CDF were measured in
the stack downstream of the control device. The average emission levels of
CDD/CDF over three runs were approximately 7.7 ng/dscm at 7 percent 02-33
Sixty-two 15-minute averages of the CO emissions from the Duluth FBC indicated
levels between 1 and 31 ppm. Controlled particulate emissions levels were
0.0048 gr/dscf (11 mg/dscm).
3.8.1.2 LaCrosse. Wisconsin
Diagnostic and compliance tests were conducted at the Northern States
Power French Island Facility Unit #1 at Lacrosse. WI in May 1988. All testing
was conducted at the stack, downstream of the electrified granular bed filter.
The unit combusts a mixture of fluff RDF [185 tpd (168 Mg/day)] and wood chips
[175 tpd (159 Mg/day)]. The average CDD/CDF emissions level observed in three
runs was 14.3 ng/dscm.34 CO emissions at NSP averaged 275 ppm over three
runs. The total particulate concentration at the stack averaged 0.0184
gr/dscf (42 mg/dscm) over three runs. The gravel bed scrubber reportedly
operates at 325°F (163°C).
3.8.1.3 Sundsvall . Sweden
Stack emissions test data are reported by Gotaverken for a CFB facility
in Sundsvall. Sweden co-firing peat. wood, and tires with RDF.35 Weight
percentages of each of the fuels are not specified in the test reports. The
average reported CDD/CDF concentration was 392 ng/Nm3. Carbon monoxide
emission data from Sundsvall indicate that the units can maintain
concentrations below 100 ppmv.
3-47
-------
3.8.2 Model Plant Baseline Emissions
The data from the Duluth and the LaCrosse FBCs are used to establish
baseline CDD/CDF emissions for the bubbling bed model plant (20 ng/dscm). The
emissions data from Sundsvall are used to estimate baseline emission for the
CFB models (400 ng/dscm). Both models were assumed to have CO emissions of 50
ppm. Uncontrolled PM emissions are assumed to be equivalent to PM emissions
from RDF spreader stokers [4 gr/dscf (9200 mg/dscm)].
3-48
-------
4.0 REFERENCES
1. Radian Corporation. "Municipal Waste Combustors - Background
Information for Proposed Standards: lll(b) Model Plant Description and
Cost of Control." EPA-450/3-89-27c. August 1989.
2. Assessment of Municipal Waste Combustor Emissions Under the Clean Air
Act. U.S. EPA Advance Notice of Proposed Rulemaking. 52 FR 25399. July
7. 1987.
3. "Municipal Waste Combustion Study: Report to Congress." EPA/530-SW-87-
021a. May 1987.
4. "Municipal Waste Combustion Study: Combustion Control of MSW Combustors
to Minimize Emission of Trace Organics." EPA/530-SW-87-021c. May 1987.
5. Environment Canada. NITEP. "Two Stage Combustion." EPS 3/UP/l.
September 1986.
6. New York State Energy Research and Development Authority. "Results of
the Combustion and Emissions Research Project at the Vicon Incinerator
Facility in Pittsfield. MA." June 1987.
7. Radian Corporation. "Municipal Waste Combustion Multipollutant Study -
Summary Report." North Andover RESCO, North Andover, MA. EMB Report
No. 86-MIN-02A. March 1988.
8. Entropy Environmentalists. "Stationary Source Sampling Report -
Pinellas County Resource Recovery Facility." St. Petersburg, FL.
February and March 1987.
9. Stieglitz, Vogg. "New Aspects of PCDD/PCDF Formation in Incinerator
Processes." Presented at NITEP Conference on Municipal Waste
Incineration. Montreal, Quebec. October 1-2, 1987.
10. Hagenmaier. et al . "Catalytic Effects of Fly Ash from Waste
Incineration Facilities on the Formation and Decomposition of
Polychlorinated Dibenzo-p-dioxins and Polychlorinated Dibenzofurans."
Environmental Science and Technology. November 11. 1987, Vol. 21. 1080-
1084.
11. New York State Energy Research and Development Authority. "Results from
the Analysis of MSW Incinerator Testing at Peekskill. NY." DCN:88-233-
012-21. August 1988.
12. Radian Corporation, "lll(b) MWC Profile Memorandum." Submitted to EPA
September 9. 1988.
13. Entropy Environmentalists. "Emissions Test Report, Municipal Waste
Combustion Study, Wheelabrator Millbury. Inc." Prepared for the U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina.
EPA/EMB Report No. 88-MIN-07. July 1988.
14. "Municipal Waste Combustion Study: Emissions Data Base." EPA/530-SW-87-
021b. May 1987.
4-1
-------
15. Radian Corporation. "Emission Test Report - Marion County Solid Waste-
to-Energy Facility." Brooks. Oregon. EMB Report No. 86-MIN-03.
September 1987.
16. "Emissions Test Results for the PCDD/PCOF Internal Standards Recovery
Study Field Test; Runs 1. 2, 3, 4. 5. 13. and 14." Memorandum from
Michael A. Vancil. Radian Corporation, to C.E. Riley. EPA. July 24.
1987.
17. Ogden Martin. "Environmental Test Report - Alexandria/Arlington
Resource Recovery Facility Units 1, 2. and 3." March 9. 1988.
18. Wheless. E. "Air Emission Testing at the Commerce Refuse to Energy
Facility." Presented at NITEP Conference on Municipal Waste
Incineration. Montreal. Quebec. October 1-2, 1987.
19. Linz. D., Gas Research Institute; Fleming. D.. Khinkis. M.. Abbasi, H.,
Institute of Gas Technology: Penterson, C., Riley Stoker Corporation.
"Emissions Reduction From MSW Combustion Systems Using Natural Gas."
Paper presented at International Conference on Municipal Waste
Combustion. Hollywood. FL, April 11-14. 1989.
20. Rasmussen. M. "Emission Test at a Danish Energy From Waste Plant."
Provided to EER Corporation by Volund USA Ltd. on May 19. 1988.
21. EER Corporation and Radian Corporation. "Municipal Waste Combustors -
Background and Information for Proposed Guidelines for Existing
Facilities." EPA-450/3-89-27e. August 1989.
22. Radian. Minutes from December 10, 1987 meeting between Westinghouse,
EPA. EER. and Radian at N.C. Mutual Building, Durham. NC.
23. Beachler. D., Pompelia. D.M.. and Weldon. J. "Bay County. Florida Waste-
to-Energy Facility Air Emission Tests." Presented at NITEP Conference
on Municipal Waste Incinerator, Montreal, Quebec. October 1-2, 1987.
24. Roy F. Weston, Inc. "Compliance Test Results - Pigeon Point, DE Energy
Generating Facility." January 1988.
25. Response to Clean Air Act Section 114 Information Questionnaire.
Results of Non-criteria Pollutant Testing Performed at Pope-Douglas
Waste to Energy Facility. July 1987. Provided to EPA on May 9, 1988.
26. Telecon. D. Scales. Consumat Systems and P. Schindler, EER Corporation.
June 1. 1988.
27. New York DEC. "Phase I Resource Recovery Facility Emission
Characterization Study - Overview Report." May 1987.
28. Response to Clean Air Act Section 114 Information Questionnaire.
Results of Compliance Testing Performed at Windham Energy facility.
Willamantic. CT. May 18-19, 1982. Provided to EPA on July 11. 1988.
4-2
-------
29. "Municipal Waste Combustion, Multi-Pollutant Study, Emission Test
Report, Maine Energy Recovery Company, Refuse-Derived Fuel Facility,
Biddeford, Maine. Volume I, Summary of Results." EPA-600/8-89-064a.
July 1989.
30. Barsin, J.A., Bloomer, T.M., Gonyeau. J.A., and P.K. Graika. "Initial
Operating Results of Coal-Fired Steam Generators Converted to 100%
Refuse-Derived-Fuel." Presented at the American Flame Research
Committee 1987 International Symposium of Hazardous. Municipal, and
Other Wastes. Palm Springs, CA. November 2-4. 1987.
31. Interpoll Laboratories. "NSP Red Wing RDF Plant - Results of March 1988
Compliance Test on Boiler No. 2." May 10. 1988.
32. Nelson, L.P., Energy and Environmental Research Corporation. "Municipal
Waste Combustion Assessment: Fluidized Bed Combustion." EPA-600/8-89-
061. July 1989.
33. Interpoll Laboratories. "Results of the November 3-6, 1987 Performance
Test on the No. 2 RDF and Sludge Incinerator at the WLSSD Plant in
Duluth. Minnesota." Interpoll Report No. 7-2443. April 25. 1988.
34. Clean Air Engineering. "Results of Diagnostic and Compliance Testing at
NSP French Island Generating Facility Conducted May 17-19. 1988." July
1988.
35. Kullendorf, A., Oscarsson, B., and Rollan, C. "Gotaverken CFB Boiler:
An Environmentally Safe Solution to Our Waste Disposal Crisis." Fourth
Solid Waste Management and Materials Conference. New York. January
1988.
4-3
-------
TECHNICAL REPORT DATA
rrad JntiniCtiom on "><• rrtrnr bf/fr (
EPA-600/8-89_i057
T IT L I ANt< Vt-l' T i » i. I
—i
Municipal Waste Combustion Assessment: Combustion ' .,»•<"•>
Control at New Facilities
August 1989
.fc »•' f **' O H »v . t* i, ( *- (, (• f
P.J. Schindler
9 PS ff O«MING OKGANIjATlOK *AM€ AMD AOO"CSS
Energy and Environmental Research Corporation
3622 Lyckan Parkway, Suite 5006
Durham. NC 27707
68-03-3365
1 3. S»-ONSO«ING AGtNCY NAUC Af
-------